3.1. DOC Release is Species-Dependent and Varies with Warming Exposure Times
Warming triggered a 165% rise in M. aeruginosa DOC release rates, which could represent a transition to a new metabolic state (Fernández-González et al. 2020; Moreno et al. 2023) with greater carbon export (figure 1a). On the other hand, Cyclotella sp. DOC release rates almost doubled, but only in non-acclimated strains (figure 1b). This might suggest a photosynthetic overflow as a short-term response to temperature rise (Thornton 2014), but as warming persists (as in the acclimated strains), trade-offs may come into play and the DOC release rates decrease. These findings are in line with what is known about the ecology of each group: diatoms thrive in colder environments and are particularly abundant in polar waters, while cyanobacteria prefer warmer conditions. This is linked to the contrasting effects of warming on the photosynthetic capacity of these two groups (Staehr and Birkeland 2006; Tan 2011; Chen 2015; Miettinen 2018). This corroborates the idea that phytoplankton DOC release increases as a function of temperature (Thornton 2014; Lønborg et al. 2020), at least in the short term. Also, it reiterates the importance of considering acclimation procedures, as well as their duration, in studies that are interested in evaluating this and other responses of phytoplankton to temperature (Staehr and Sand-Jensen 2006).
Figure 1 DOC release rates (slopes from the linear regression models) for each treatment. Distinct letters represent significant differences between strains (p < 0,05) from higher to lower values in alphabetical order, given by a Kruskal-Wallis test (n = 3000). An outlier in Cyclotella sp. control strain with value of ~ 2.5 is omitted for improved readability. Slope values here presented are a result of a resampling by bootstrap (1000 iterations)
3.2. Warming Affects FDOM Composition and Bioavailability.
The PARAFAC 4-component model (Fig. 2) was cross-referenced with the 267 public models of fluorescent components in the OpenFluor database, which returned 11, 29, 12 and 4 matching components for the components 1 to 4, respectively. The description given by this analysis, together with the peak classification (Stedmon et al. 2003; Coble 2007), is shown on Table 1. Interestingly, a quarter of the corresponding components are from studies conducted in polar regions, where DOM is known to be primarily autochthonous, once vascular plants are absent and algal DOM becomes prevalent (Pointing et al. 2015; Berggren et al. 2020).
Table 1
Description of the four components according to peak classification and OpenFluor cross-referencing.
|
λex max. (nm)
|
λem max. (nm)
|
Literature peaks correspondence
|
OpenFluor component matches description
|
C1
|
270
|
348
|
T (Tryptophan-like)
|
DL-tryptophan and indoles (Wünsch et al. 2015), protein-like, from microbes, algae and/or bacteria (Cory and Mcknight, 2005; Lambert et al. 2017) protein-bound or free amino acids (DERRIEN et al. 2019).
|
C2
|
275
|
300
|
B (Tyrosine-like)
|
L-tyrosine and p-cresol (Wünsch et al. 2015), tannin-like and protein-like (D’Andrilli et al. 2017), freshly produced or freshly degraded proteinaceous material (D’Andrilli et al. 2019), protein-bound or free (Chen et al. 2017).
|
C3
|
< 240(350)
|
468
|
A + C (Fulvic and humic-like)
|
Terrestrial humic-like (Dainard and Guéguen 2013; Williams et al. 2013), microbially processed (Osburn et al. 2012), photo-refractory (Kida et al. 2019).
|
C4
|
< 240(315)
|
396
|
A + M (Marine/Microbial humic-like)
|
Microbial humic-like (Lapierre and Del Giorgio 2014; Wang et al. 2020), present in algae cultures (Søndergaard et al. 2011).
|
Among the 4 fluorophores detected in this set of axenic phytoplankton cultures, three are more commonly recognized as microbially-derived (components 1, 2 and 4) and one is rarely attributed to this source (component 3). This highlights a relevant aspect for interpreting the occurrence of humic substances in natural waters: many studies indicate that such compounds are also produced in the absence of terrestrial inputs, and thus the assumption of an allochthonous origin must be made with caution (Vines and Terry 2020; Amaral et al. 2021). Also related to this finding are the low FI values found in the exponential and stationary samples of
Cyclotella sp. (1.34–1.36, see figure S4), below the threshold for DOM of mixed sources (~ 1.4) and approaching the reference value for terrestrial DOM (~ 1.2) (Gabor et al. 2014; Hansen et al. 2016; Melo et al. 2020).Fluorescent DOMp concentration and composition exhibited shifts in response to both warming and acclimation time (Fig. 3 and figure S5). In
M. aeruginosa cultures, contrasts were only observed in the stationary growth stage. The component 2 mean relative fluorescence (%C2) is significantly higher at 28°C than at 24°C. Conversely, the component 3 mean relative fluorescence (%C3) is higher in the control than in the acclimated cultures, both with no significant difference to non-acclimated cultures, which show intermediate values. The humic-like compounds associated to C3 are considered less bioavailable, while the protein-like compounds associated to C2 are highly labile (Cory and Kaplan 2012; Yan et al. 2022), thus the observed changes indicate higher exudate bioavailability.
Changes in HIX values for stationary M. aeruginosa samples endorse these results. In control cultures, HIX values (2.54) are greater than in acclimated cultures (1.77), with non-acclimated cultures showing an intermediate value (2.22) (see figure S4). High HIX values are indicative of further polycondensation and reduced hydrogen/carbon ratio of organic molecules, characteristics of more recalcitrant DOM (Fellman et al. 2010; Gabor et al. 2014; Yan et al. 2022), so the decrease in HIX values with temperature rise also points to an increase in bioavailability in response to warming.
Now looking at absolute fluorescence data for stationary M. aeruginosa, an almost twice fold rise in all components happened in acclimated cultures. That could indicate an increase in the proportion of fluorophores in the DOMp when this cyanobacteria becomes acclimated to a higher temperature (see figure S5 for further comparisons).
In turn, trends in FDOM from Cyclotella sp. cultures suggest a less labile exudate in warmer conditions. Both C1 in the stationary stage and %C1 in the exponential stage decreased with higher temperature and warming exposure time. Simultaneously, both %C4 in the stationary stage and C3 in the exponential stage tended to increase. Considering the protein-like and more labile (Cory and Kaplan 2012; Yan et al. 2022) characteristics of C1 and the humic-like and possibly less labile features (Cory and Kaplan 2012; Yan et al. 2022) of C3 and C4, these changes represent a reduction in bioavailability of FDOM from Cyclotella sp. exudates under warmer conditions. Again, HIX values support this interpretation, at least partially: a higher mean value is observed for non-acclimated cultures (0,81), significantly greater than the value for control cultures (0,55) but not than the values for acclimated ones (0,66) (figure S4).
3.3. Warming affects FDOM composition and concentration as much as growth state change
When evaluating whether changes in growth stage or temperature were the main drivers of variation in fluorescent DOMp, an overall pattern of similar influence emerged (Fig. 4). In almost all cases, fluorescent DOMp concentration and composition was determined by the joint action of growth state change and temperature. The only exception in the general pattern was observed for FDOM concentration in M. aeruginosa cultures. This could be due to the sharp variation in cell density observed from the exponential to the stationary stage, approximately threefold, which represents a strong contrast to the nearly constant values observed in Cyclotella sp. cultures (Figure S6).
Here it is worth noting that the present investigation did not set out to estimate FDOM degradation by phytoplankton, but once the concentration of all components has increased over time, it is assumed that the processes of production, release and transformation outweigh those of absorption and degradation. Also, it is fundamental to frame into context the findings reported here, as they come from relatively short assays with isolated species under very controlled conditions – also, the analytical window of fluorescence spectroscopy is limited and there is a persistent challenge to make progress towards a molecular comprehension of DOM (Murphy et al. 2010; Stubbins et al. 2014; Catalán et al. 2020). Thus, long-term phytoplankton evolutionary adaptations, ecological interactions with other microbial communities – especially bacteria (Amin et al. 2015; Johansson et al. 2019; Shibl et al. 2020) – and the influence of synergistic effects with other environmental shifts associated with climate change (such as eutrophication and rise in atmospheric CO2 levels) were beyond the scope of this work, and extrapolations must be made wisely.
Nonetheless, the outcomes here reported allow us to set some expectations on the response of aquatic microbiomes to climate change. The competitive advantages of phytoplankton taxa with better fitness to warmer conditions can influence microbiome species composition (Bagatini et al. 2014), which added to metabolic changes that affect the release and bioavailability of DOMp, even at small scales, can trigger disruptions in microbial communities and in the structure of the (osmo)trophic web, impacting the speed of efficiency and intensity of carbon flows in aquatic trophic webs (Sarmento and Gasol 2012; Landa et al. 2016; Tisserand et al. 2020), especially in the so-called “microbial loop”.