Environmental factors controlling the distributions of Botryococcus braunii (A, B and L) biomarkers in a subtropical freshwater wetland

Here we report the molecular biomarker co-occurrence of three different races of Botryococcus braunii (B. braunii) in the freshwater wetland ecosystem of the Florida Everglades, USA. Thespecific biomarkers include C32–C34 botryococcenes for race B, C27–C32 n-alkadienes and n-alkatrienes for race A, and lycopadiene for race L. The n-alkadienes and n-alkatrienes were present up to 3.1 and 69.5 µg/g dry weight (dw), while lycopadiene was detected in lower amounts up to 3.0 and 1.5 µg/g dw in periphyton and floc samples, respectively. Nutrient concentrations (P and N) did not significantly correlate with the abundances of these compounds. In contrast, n-alkadienes and n-alkatrienes were present in wider diversity and higher abundance in the floc from slough (deeper water and longer hydroperiod) than ridge (shallower water and shorter hydroperiod) locations. n-Alkadienes, n-alkatrienes, and lycopadiene, showed lower δ13C values from −40.0 to −35.5‰, suggesting that the source organisms B. braunii at least partially utilize recycled CO2 (13C depleted) produced from OM respiration rather than atmospheric CO2 (13C enriched) as the major carbon sources.


Experimental Methods
Sampling locations. The sampling sites for this study feature a gradient of nutrient and hydroperiod in the Everglades ( Fig. 1; Table 1; http://fcelter.fiu.edu/data) 19 . Briefly, Water Conservation Area 3 (WCA3) is located TerraMetrics) and then drafted by using software CorelDRAW (Graphics Suite × 6, source ID: 017002) (http:// www.coreldraw.com/en/product/graphic-design-software). The sampling sites were located by using Global Positioning System (GPS). to the north of Everglades National Park (ENP),and has the longest hydroperiod and highest nutrient (P and N) levels among all freshwater sites. Sites SRS1 to SRS3 are located in the Shark River Slough within ENP, characterized by intermediate hydroperiod and nutrient levels. Sites TSPh2-4 are located in the Taylor Slough within ENP, with lower hydroperiod and nutrient levels ( Table 1). All these sites (SRS1-3 and TSPh2-4) are characterized by diverse aquatic vegetation and microbial communities 12 .
Sample collection and preparation. Periphyton and floc (regardless of ridge and slough environments) were collected from various locations across the Florida Everglades (Table 2). Additional floc samples were sampled from both ridge and slough environments, and during different times of the year from 2012 to 2014 within WCA3. Both submerged and floating periphyton were placed into Ziploc bags. Floc, surface soils were sampled following the procedures as described previously 20,21 . Both leaves and roots (when applicable) of the dominant plants such as Nymphaeaceae, Utricularia sp., Chara sp., Cladium sp., Eleocharis sp., Typha domingensis P., and Typha latifolia L. were randomly sampled from different locations within ENP and WCA3 22 . All samples were kept cool on ice during transport to the laboratory. The samples were transferred into pre-combusted glass jars and stored at −20 °C until further analysis. All samples were processed and analyzed as described previously 21 . Briefly, they were freeze-dried at −50 °C, then shredded and sieved through a 32 mesh (500 µm) sieve to remove coarse material. Samples (1-3 g dry weight) were subjected to ultrasonic extraction three times (0.5 h each) with dichloromethane (DCM) (Optima, Fisher, USA) as solvent. Total extracts were concentrated and then fractionated by adsorption chromatography over silica gel. The aliphatic fraction and aromatic hydrocarbon fraction were eluted sequentially using n-hexane and hexane: toluene (3:1, v:v), respectively. Ziploc bags used for sampling were washed with n-hexane and the extracts were employed as control blanks and randomly analyzed to exclude external contamination.
Bulk parameter analysis. Total nitrogen (TN) was measured by the high-temperature dry combustion method using a Carlo-Erba NA-1500 CNS Analyzer. Total P was analyzed with a Technicon Auto Analyzer II System (Pulse Instruments Ltd.), according to the standard method for orthophosphate P (EPA method 365.1). Bulk δ 13 C values were also determined for floc samples using standard elemental analyzer isotope ratio mass spectrometer (EA-IRMS) procedures 23 , and reported with respect to the Vienna PeeDee Belemnite (VPDB) standard for carbon. Precision of the δ 13 C measurements was ±0.10‰.

Gas chromatography-mass spectrometry (GC-MS). GC-MS analyses were performed on a
Hewlett-Packard 6890 GC fitted with a Rtx-1 capillary column (30 m, 0.25 mm ID, Restek, USA) interfaced to a HP 5973 MSD. Compounds were quantified by squalane as the internal standard, assuming a similar response factor. Kovats retention indexes (RI) 24 were calculated following the formula: RI = 100 × (R x −R n )/(R n+1 − R n ) + 100n, where x denotes the compound of interest, R denotes the GC retention time, and n and n + 1 denote the carbon number for the nearest n-alkane and (n + 1)-alkane eluting before and after x, respectively on the GC. The identification of individual compounds was based on the comparison with published mass spectra and interpretation of the mass spectra 1 .
Gas chromatography-isotope ratio mass spectrometry (GC-IRMS). The δ 13 C values of individual n-alkadienes, n-alkatrienes and lycopadiene were measured using a GC-IRMS system, which consists of a HP 6890 GC equipped with a Rtx-1 fused silica capillary column (30 m, 0.25 mm ID), a combustion interface (Finnigan GC combustion IV), and a Finnigan MAT delta Plus mass spectrometer 21 . Between every three    samples, three standard mixes containing squalane and C 17 n-alkane (different concentrations as 30 ng/µL, 100 ng/ µL and 500 ng/µL, with known δ 13 C values for each compound) were analyzed to check instrument performance and also for correction purposes. A known amount of squalane was used as an internal standard. The δ 13 C values are given in the per mil (‰) notation relative to the Vienna PeeDee Belemnite (VPDB) standard. The reproducibility of the GC-IRMS system was <0.5‰ for both standards and repeat analyses of selected samples (n = 3). Due to the co-elution of a few n-alkadiene or n-alkatriene isomers, and the relative lower concentration for some specific non-dominant isomers, only compounds present in sufficient quantities (intensity above 1000 mVs) could be accurately determined for reliable δ 13 C values. Average values were reported if more than one δ 13 C value was measured for isomers with the same carbon atom numbers.

Data analysis. Environmental data across multiple locations was obtained from the Florida Coastal
Everglades Long Term Ecological Research database (FCE-LTER; http://fcelter.fiu.edu/) and used for comparison with the abundance of the biomarker compounds (botryococcenes, n-alkadienes, n-alkatrienes, and lycopadiene). Statistical analyses were performed using SPSS version 13.0 for Windows. Outliers were tested using the two-sided Grubbs test (P < 0.05). Significant correlations (P < 0.05) between floc physicochemical parameters and the biomarker compounds were determined using Pearson correlation. Significant differences between means of different groups of data were compared using the unpaired t-test (two-tailed, unequal variance).

Identification of n-alkadienes, n-alkatrienes and lycopadiene. GC-MS analysis of the n-hexane
eluted fraction from various periphyton and floc sample extracts showed the presence of n-alkadienes, n-alkatrienes and one lycopadiene ( Fig. 2; Table 2). A total of 11 C 27 to C 32 n-alkadiene and n-alkatriene isomers eluting between the C 26 to C 32 n-alkanes were tentatively identified and their Kovats indexes are also given (Figs 2a-f; 3a-k; Table 2). These compounds all exhibit a terminal double bond and one or two mid-chain unsaturations with both Z and E stereochemistry 1 . The mid-chain double bond positions could be further identified based on dimethyl disulfide adducts experiments 25 , but this is not pursed in this present study. Generally, no carbon number predominance was found for these n-alkadienes and n-alkatrienes. Similar no odd or even carbon chain predominance was observed in C 37 -C 43 mono-, di-and (to a lesser extent) tri-unsaturated n-alkenes reported in lacustrine sediments 26 . Lycopa-14,18-diene was identified based on its retention time and mass spectrum match with that in the literature (Fig. 3) 1,9 . Another lycopadiene isomer with lower abundance was also identified (Fig. 3). A series of botryococcenes with 32 to 35 carbon atoms are detected in most periphyton and floc samples and elute between C 26 to C 29 n-alkanes, in agreement with a previous report (Fig. 2) 18 .

Spatial distribution of n-alkadienes, n-alkatrienes and lycopadiene.
A higher molecular diversity of n-alkadienes and n-alkatrienes was detected in floc compared to periphyton (Table 2). Specifically, only C 29 and C 31 n-alkadienes, and C 30 n-alkatriene were found in periphyton samples, while C 27 -C 31 n-alkatrienes were present in floc samples. Lycopadiene occurred in most of the periphyton samples, but rarely in floc samples. Floc samples (n = 86) from both ridge and slough locales within the WCA3 area and floc samples (n = 12) from SRS2 and TSPh2 were analyzed. The N and P (nitrogen and phosphorus) concentration of these floc samples were 9.7-46.2 mg/g dw and 73-884 µg/g dw, respectively. The total concentration of n-alkadienes and n-alkatrienes of these floc samples ranged from 135 to 6953 ng/g dw. Surprisingly, the abundance of the C 29 n-alkatriene could be up to 2 times above that of the C 29 n-alkane in the same sample (Fig. 2), which is in contrast with previous reports that n-alkadienes and n-alkatrienes usually show much lower abundances than the odd numbered n-alkane homologues 1 . No significant correlations were observed (P > 0.05) between nutrient concentrations and the concentrations of each compound group or the total concentrations in floc (Fig. 4). In addition, no significant correlations were observed between surface water nitrogen and phosphorus concentrations, and abundances of n-alkadienes and n-alkatrienes among different locations across the freshwater wetland. Floc samples from ridge (n = 19) and slough (n = 12) environments (within 5 m distance) were analyzed from multiple years (2012 to 2014) within the WCA3 area. The concentrations of n-alkadienes and n-alkatrienes in the slough floc samples ranged from 2.0 to 69.5 µg/g dw (average as 13.9 µg/g dw). In contrast, the concentrations of n-alkadienes and n-alkatrienes in the ridge floc samples ranged from 0 to 6.6 µg/g dw (average as 1.6 µg/g dw; Fig. 4). The concentrations of n-alkadienes and n-alkatrienes were significantly higher in the slough than the ridge floc (unpaired student t-test, two tailed, P < 0.01). In addition, 8 transects were analyzed from slough to the ridge environments (n = 32) and obvious concentration decrease trends were observed (Fig. 4).

Compound specific carbon isotopes of n-alkadienes, n-alkatrienes and lycopadiene.
Compound specific stable carbon isotope analysis was performed on the dominant n-alkadienes, n-alkatrienes and lycopadiene. Due to incomplete GC resolution of some n-alkadiene or n-alkatriene isomers with the same carbon number, the δ 13 C values are reported as averages for those compounds (mixtures of Z and E isomers). Significantly lower δ 13 C values were observed for the n-alkadienes, n-alkatrienes and lycopadiene ( Table 2) than those of n-alkanes (−32.7 ± 1.8‰) and bulk samples (−30.7 ± 1.4‰). No significant differences in the averaged δ 13 C values were observed between n-alkadienes and n-alkatrienes, whereas the averaged δ 13 C values of the n-alkadienes and n-alkatrienes were lower than those of lycopadiene (Table 3).

Co-occurrence of B. braunii (A, B, L) indicated by n-alkadienes, n-alkatrienes, and lycopadiene in the Everglades.
Lycopadiene has been reported as a specific biomarker for race L of B. braunii 4 , while botryococcenes have been suggested to derive from race B of B. braunii in the Everglades 18 . n-Alkadienes and n-alkatrienes were not detected in floc and surface soil at the mangrove-dominated estuarine locations 27 , nor in the leaves or roots of dominant plant species across the Everglades ecosystem 22,28 . n-Alkadienes and n-alkatrienes have been reported in insect wax lipids, but they usually cover higher carbon chain lengths up to C 39 29 . Odd numbered carbon (poly) unsaturated n-alkenes in the range C 23 -C 31 have been characterized in the chlorophyte Chlorella emersonii 30 , the diatom Rhizosolenia setigera 31 , and two marine eustigmatophytes 32 . C 29 , C 31 and C 33 alkenes with one to four double bonds are also produced by some haptophytes, such as Emiliania huxleyi, Isochrysis galbana, Gephyrocapsa oceanica and Chrysotila lamellosa [33][34][35][36] . However, the n-alkadienes and n-alkatrienes detected in this study are all from only the freshwater wetland locations, and thus should not be derived from haptophytes. No significant correlations (concentration based) were observed between n-alkadienes, n-alkatrienes, and the C 20 HBI (highly branched isoprenoid) across the whole sample set (n = 98, P > 0.05), excluding cyanobacteria as the major source of n-alkadienes, n-alkatrienes detected 20,37 . Therefore, we suggest that the n-alkadienes and n-alkatrienes detected in this study likely derive from the A race of B. braunii 5 . Combining the botryococcenes and lycopadiene produced by the B and L races of B. braunii, the co-occurrence of three races (A, B and L) of B. braunii seems possible. No significant correlations exist among the abundances of biomarkers of different races of B. braunii, which could be caused by: (1) variations in the populations of each race of B. braunii across our study area, (2) differences in the hydrocarbon concentration in each race, and (3) different physiological states for each race 4 . Similar results have also been observed in another study 1 . Mixtures of cis n-alkadiene and triene(s) or cis/trans dienes (without triene) covering the carbon chain range from C 25 to C 31 have been characterized in the A race of B. braunii 3,6,7,38,39 . Cis-dienes have been suggested to be produced via an elongation-decarboxylation mechanism with oleic acid as the direct precursor 40 . The L race of B. braunii can produce lycopadiene 5 . Recently, the microfossils of B. braunii have been observed in soil cores of tree islands, and floc at WCA3 area in the Everglades 16 , providing evidence for the existence of B. braunii in this ecosystem. Unfortunately, no specific race of B. braunii was described in previous studies.
Lower δ 13 C values were obtained for n-alkadienes, n-alkatrienes, and lycopadiene, which are similar to those observed previously for botryococenes 18 , suggesting that these compounds are likely produced by B. braunii utilizing at least partially recycled ( 13 C depleted) CO 2 from organic matter degradation as their carbon sources rather than atmospheric ( 13 C enriched) CO 2 21 . Similar lower δ 13 C values have also been observed for C 20 HBIs (for cyanobacteria) in the freshwater Everglades periphyton and floc. Although the δ 13 C values of the n-alkadienes and n-alkatrienes, to our best knowledge, have not yet been reported, the δ 13 C values for the biomarkers of B. braunii are diverse 41 . The δ 13 C values of botryococcenes (or botryococcanes) and lycopadiene-derived compounds are reported from −37.4‰ to −10.6‰ and −29.0‰ to −21.0‰, respectively 9,18,41,42 . Even though, Boreham et al. 43 stated that the large range of δ 13 C values may not be fully expressed due to differences of internal diffusion rates of CO 2 , this wide range of δ 13 C values is at present not clear 43 .

Environmental controls of B. braunii biomarkers and their implications in the Everglades.
Although B. braunii is known to be sensitive to environmental changes 42 , and botryococcenes have been suggested to be applied as a proxy for eutrophication, the lack of correlations between nutrients and n-alkadienes and n-alkatrienes suggest that they seem not to be indicators for eutrophic conditions in this freshwater wetland. Actually, Botryococcus was also not suggested to directly reflect nutrient status of waters in the Everglades 17 .
In contrast to nutrients, hydroperiod seems to be one of the controlling factors for the distribution of n-alkadienes and n-alkatrienes. Significant higher abundances of n-alkadienes and n-alkatrienes were observed in the ridge than slough floc, which could be explained by the following reasons: (1) the A race of B. braunii has the ability to float due to its high lipid concentrations 44 , which leads to its enrichment in the slough environment 17 ; and (2) more diagenetic degradation of n-alkadienes and n-alkatrienes occurs in the ridge environment due to stronger oxidation. In this study, n-alkadienes and n-alkatrienes were only observed in the floc at locations SRS2 and WCA3, in agreement with higher concentrations of botryococcenes reported at these two sites 18 . This could be mainly attributed to longer hydroperiod (WCA3 and SRS2), and lower water flow velocities (WCA3), resulting in reduced floc transport ( Table 1). The longer hydroperiod could induce sub-oxic or anoxic conditions in the floc layer, and thus decrease carbon mineralization rates. However, other factors may also contribute to the concentration difference among different sites, such as the composition of periphyton and these require further investigations (Table 1) 45 .
n-Alkadienes and n-alkatrienes were only detected in WCA3 slough soils/sediments (Fig. 3), but not in all SRS and TSPh sites, which could likely be due to: (1) a more complex microbial composition in periphyton and floc compared to soils 13,14 , (2) limited incorporation of periphyton and floc into soils, or (3) early diagenetic reworking or microbial degradation of these compounds by heterotrophs such as bacteria and fungi 46 . Several sulfur-containing compounds and two thiophenes both with 20 carbon atoms (3-methyl-2-(3′,7′,11′trimethyldodecyl)thiophene and 3-(4′,8′,12′-trimethyltridecyl)thiophene were detected in most of the floc samples (data not shown), suggesting anoxic or sub-oxic conditions. If early diagenetic reduction of the unsaturations is one of the factors accounting for the absence of n-alkadienes and n-alkatrienes in all deeper soils/ sediments, part of the C 27 -C 33 n-alkanes detected in sediments of the Florida wetland could also be derived from the n-alkadienes and n-alkatirenes 47 . However, further investigation is needed. Lycopadiene was not detected in surface and deeper soils, likely due to: (1) a much lower amount of lycopadiene produced, or (2) diagenetic transformation of lycopadiene into higher molecular weight compounds 47 . However, lycopadiene was reported in a few studies including freshwater lake sediments from the Holocene 1 and an oil shale from the Pliocene 47 . In addition, a monoaromatic lycopane derivative was reported from the Messel oil shale 9 , and kerogen fractions of samples from oil shale layer 4 in the Eocene Huadian Formation, NE China 48 . Adam et al. 9 proposed that this compound could be a specific biomarker for race L of B. braunii in sediments deposited under freshwater and/or brackish conditions 9 . Though analyzing a Holocene freshwater lake sediment core, Zhang et al. 1 suggested that n-alkadienes, botryococcenes and lycopadienes can survive in oxic sediments for several decades, and the down core variation in these lipids likely reflects changes in environmental conditions either favoring the bloom or near-extinction of B. braunii 1 . However, this present study shows that botryococcenes were widely detected, and ScIEntIfIc REpORTS | (2018) 8:8626 | DOI:10.1038/s41598-018-26900-9 n-alkadienes and n-alkatrienes were rarely present, while lycopadiene was not detected in the surface and deeper soils of this subtropical freshwater wetland. This possibly suggests either their general rapid diagenetic reworking, or more likely to a recent increasing contribution of Botryococcus-derived organic matter input in the Everglades. Further studies are needed to address the other factors controlling the distribution of these biomarkers in order to better use them as indicators of the B. braunii community structure in the Everglades.

Conclusions
A series of long chain n-alkadienes and n-alkatrienes, botryococcenes and a lycopadiene were detected in periphyton and floc across the freshwater wetlands of the Florida Everglades, USA, suggesting the co-existence of all three races of the green alga B. braunii (A, B and L). Similar low δ 13 C values were observed for n-alkadienes and n-alkatrienes, lycopadiene and botryococcenes, suggesting that the source organisms (B. braunii) at least partially utilize recycled CO 2 produced from respired organic matter rather than atmospheric CO 2 as the carbon sources. The concentrations and molecular distributions of these compounds were shown to decrease from floc to periphyton. n-Alkadienes, n-alkatrienes and lycopadiene were not found in soils, suggesting a recent contribution of these compounds likely due to the blooming of B. braunii.
The abundance of these compounds does not correlate with both bulk N and P concentrations in floc samples or surface water, suggesting that nutrients may not be the controlling factors for the distributions of these compounds in this ecosystem. In contrast, slough floc contains significantly higher amounts of n-alkadienes and n-alkatrienes than ridge floc. Thus, hydroperiod could be one of the controlling factors for the abundances of n-alkadienes and n-alkatrienes within this freshwater wetland. However, further investigation is needed to refine the application of these biomarkers asl indicators of community structure of B. braunii in the Everglades.