Temperature Stress Induces Shift From Co-Existence to Competition for Organic Carbon in Microalgae-Bacterial Photobioreactor Community - Enabling Continuous Production of Microalgal Biomass

Abstract

To better predict the consequences of environmental change on aquatic microbial ecosystems it is important to understand what enables community resilience. The mechanisms by which a microbial community maintain its overall function, for example, the cycling of carbon, when exposed to a stressor, can be explored by considering three concepts: biotic interactions, functional adaptations, and community structure. Interactions between species are traditionally considered as, e.g., mutualistic, parasitic, or neutral but are here broadly defined as either coexistence or competition, while functions relate to their metabolism (e.g., autotrophy or heterotrophy) and roles in ecosystem functioning (e.g., oxygen production, organic matter degradation). The term structure here align with species richness and diversity, where a more diverse community is though to exhibit a broader functional capacity than a less diverse community. These concepts have here been combined with ecological theories commonly used in resilience studies, i.e., adaptive cycles, panarchy, and cross-scale resilience, that describe how the status and behavior at one trophic level impact that of surrounding levels. This allows us to explore the resilience of a marine microbial community, cultivated in an outdoor photobioreactor, when exposed to a naturally occurring seasonal stress. The culture was monitored for 6weeks during which it was exposed to two different temperature regimes (21 ± 2 and 11 ± 1°C). Samples were taken for metatranscriptomic analysis, in order to assess the regulation of carbon uptake and utilization, and for amplicon (18S and 16S rRNA gene) sequencing, to characterize the community structure of both autotrophs (dominated by the green microalgae Mychonastes) and heterotrophs (associated bacterioplankton). Differential gene expression analyses suggested that community function at warm temperatures was based on concomitant utilization of inorganic and organic carbon assigned to autotrophs and heterotrophs, while at colder temperatures, the uptake of organic carbon was performed primarily by autotrophs. Upon the shift from high to low temperature, community interactions shifted from coexistence to competition for organic carbon. Network analysis indicated that the community structure showed opposite trends for autotrophs and heterotrophs in having either high or low diversity. Despite an abrupt change of temperature, the microbial community as a whole responded in a way that maintained the overall level of diversity and function within and across autotrophic and heterotrophic levels. This is in line with cross-scale resilience theory describing how ecosystems may balance functional overlaps within and functional redundancy between levels in order to be resilient to environmental change (such as temperature).

Keywords: adaptive cycles; bacteria; coexistence; community; competition; interactions; microalgae; resilience.

Figure 1

Parameters associated with the photobioreactor (PBR) during the sampling period, dates of sampling specified on top, sample names S1–S8 at the bottom. (A) Daytime temperature (°C) in PBR (closed circles), mean temperature per 24h in PBR (gray triangles), regional ambient daytime temperature (measured by SMHI; open squares). (B) Number of hours day^-1 with sunshine (direct solar radiation > 120Wm^–2, 10km south of PBR location; open circles), daytime hours/day (black triangles). (C) For S2–S8 sampling events (S1 omitted due to lack of data), left y-axis: Dry weight (DW) of biomass (gl⁻¹; open triangles) and right y-axis: ratio of flow in added mass (g) of CO₂:total dry weight (in 3200L; asterisks).

Figure 2

Canonical correspondence analysis (CCA) biplot of the eukaryote (A) and prokaryote (B) community, plotting samples with independent environmental parameters at each date of sampling (S2–S8) in the reactor: Temp – temperature in PBR (°C), DW – biomass dry weight (gl⁻¹), sunh – number of sunshine hours at day of sampling (h), light – unshaded/shaded (reduced light input by 40–60%). Dates are denoted by color, warm period by (○), and cold period by (△). Date S1 is excluded due to missing nutrient data.

Figure 3

Taxonomic affiliation of ASVs, mean of triplicates per sampling date (S1–S8), with a relative abundance > 0.1%. (A) 18S ASVs with assigned taxonomy at the genus level. (B) 16S ASVs with assigned taxonomy at the order level. Warm: 19.5 ± 0.89°C, cold: 12.4 ± 1.76°C (mean temperature per 24 h in PBR); Light: light reduction, open bar → natural light, striped bar → light reduced by 40–60% by shading.

Figure 4

Diversity measures of 16S ASVs, excluding ASVs annotated as chloroplasts. (A) Rarefaction curves, plotting sample size vs. number of ASVs. (B) Normalized rank abundance curves, plotting ASVs, ranked by abundance vs. abundance (counts). Orange – warmer period, blue – colder period.

Figure 5

Eukaryote gene expression (metatranscriptome) of mechanisms associated with carbon acquisition. Included are key enzymes representative of: carbon concentration (C-conc) and carbon fixation (C-fix), ketone body formation, hydrolases, and carbon transport. Heatmap shows the mean of triplicates per sampling date (S1–S8), of counts normalized to TPM and then square root transformed. Heatmap was made using R package pheatmap, with setting scale by “row.” Warm: 19.5 ± 0.89°C, cold: 12.4 ± 1.76°C (mean temperature per 24 h in PBR); Light: light reduction, open bar → natural light, striped bar → light reduced by 40–60% by shading. * indicate padj < 0.01 for log2 fold change of differential expression, contrasted for warm vs. cold time period.

Figure 6

Prokaryote gene expression (excluding transcripts annotated as eukaryotes in the prokaryotic dataset) of mechanisms associated with carbon acquisition, putative extracellular polysaccharide (EPS) formation, and respiration. Included are key enzymes representative of: carbon uptake (C-uptake), alternative carbon sources, carbon fixation, EPS-formation, glycolysis, pyruvate metabolism, tricarboxylic acid (TCA) cycle, and oxidative phosphorylation. Heatmap shows the mean of triplicates per sampling date (S1–S8), normalized to TPM and square-root transformed, made using R package pheatmap, with setting scale by “row.” Warm: 19.5 ± 0.89°C, cold: 12.4 ± 1.76°C (mean temperature per 24 h in PBR); Light: light reduction, open bar → natural light, striped bar → light reduced by 40–60% by shading. * indicate padj < 0.01, and ¤ padj < 0.05, for log2 fold change, differential expression, contrasted for warm vs. cold time period.

Figure 7

(A) Co-occurrence network (Pearson) of eukaryote (genus level) and prokaryote (order level) ASVs (network nodes). Only edges with weight > 0.3 were included in the plot for ease in visualization, and nodes with zero edges were excluded. Modules represent node clusters grouped based on their shared positive correlations (edges). We clustered modules into three major clusters (based on their correlation values with temperature, see Figure 8). MI (modules 1, 2, 9), MII (modules 3, 5 – spread out due to the exclusion of edges with weight < 0.3), and MIII (modules 4, 6, 7). Relative abundances of ASVs of grouped modules are displayed in (B) (MI – 18S), (C) (MI – 16S), (D) (MII – 18S), (E) (MII – 16S), and (F) (MIII – 16S). Bar plot facets indicate temperature regime in PBR, warm: S1–S4, cold: S5–S8.

Figure 8

Pearson coefficient correlation values (red – positive; blue – negative) of network analysis modules with parameters: temp – temperature (°C); light/shade – induced shade (40–60% reduction of light); sunhours (h) – number of hours day⁻¹ with sunshine; Dry weight – of biomass (gl⁻¹). Significances indicated by *p < 0.05, **p < 0.01, and ***p ≤ 0.001. Module groups (MI, MII, and MIII) were based on shared edges (Figure 7) and similarity in correlation to temperature.

Figure 9

Conceptual model of the impact in microalgae-bacteria interactions induced by temperature stress (A). Less CO₂ got incorporated (1) while microalgal excretion of organic C (OC) was utilized by bacteria (2), leading to a higher diversity of the microalgae (3) and a lower diversity of the bacterial community (4). Resulting in coexistence (5), due to the partitioning of carbon resources (6). The release from temperature stress (B) introduced more CO₂ to the system resulting in a higher accumulation of microalgal biomass (7), while less OC got excreted (8), leading to a lower diversity of the microalgae, being dominated by one species (9), while the bacterial diversity became higher (10). This resulted in competition between microalgae and bacteria for organic carbon (11), due to mixotrophic microalgal uptake of both CO₂ and OC (12). The partitioning of carbon resources is indicated by CO₂ (flue gas) and OC (autochthonously produced carbon; see Discussion for details).

Figure 10

An illustration of adaptive cycles and the concept of panarchy, used to describe the interactions between the microalgae (upper level) and bacteria (lower level), going from coexistence during the warmer (orange) temperature regime to competition during the colder (blue) temperature regime. r – growth phase, K – conservation phase, Ω – release phase, and α – adaptation phase. Remember – impact on lower level by upper level, revolt – influence from lower level on upper level.