Introduction

Cost-effective ways to produce large quantities of algal biomass are crucial to exploit the full potential of microalgae plants. Besides high energetic and financial expenses for optimal cultivation conditions, harvesting, and downstream processes, one major challenge to the sustainable production of algae biomass is maintaining biological stability in large-scale production plants (Scott et al. 2010; Kazamia et al. 2012; Wang et al. 2013; Carney and Lane 2014). Since sterile operation of axenic cultures at large scale is economically and practically not feasible, the formation of dynamic ecosystems in open ponds and closed photobioreactors (PBRs) is almost inevitable (Schenk et al. 2008; Cooper and Smith 2015). Recent studies have revealed the diversity and dynamics of associated microbiomes and complexity of interactions in artificial microalgae cultivation systems that seemed to be similar to the algae holobiont in natural systems (Lian et al. 2018; Ferro et al. 2020).

Algae are strongly associated with complex microbial communities in natural environments, and their interaction patterns have formed in a long coexisting evolutionary history. The phytoplankton–bacteria relationship represents a fundamental role in ecosystem functioning and has an immense impact on biogeochemical cycling in aquatic environments at a global scale (Azam 1998; Buchan et al. 2014). The microenvironment of the phycosphere allows the exchange of manifold metabolites and chemical compounds that can result in a wide spectrum of mutualistic, commensalistic, antagonistic, competitive, or parasitic relationships (Kazamia et al. 2016; Ramanan et al. 2016; Seymour et al. 2017; Cirri and Pohnert 2019). Such algae–bacterial relationships can be altered by abiotic factors like temperature and/or light availability (Mayers et al. 2016; Piwosz et al. 2020). The relationship of algae and associated bacterial communities has to be therefore considered as a dynamic range of subsequent states of highly complex networks influenced by changing environmental conditions (Gurung et al. 1999; Seyedsayamdost et al. 2014; Fuentes et al. 2016; Mayers et al. 2016; Cirri and Pohnert 2019).

Some studies conclude that algae–bacterial interactions are algae species-specific (Krohn-Molt et al. 2017; Behringer et al. 2018; Crenn et al. 2018; Koedooder et al. 2019; Mönnich et al. 2020). However, other studies suggest that these interactions were mainly based on bacterial functional traits (Ramanan et al. 2015; Geng et al. 2016a). Bacterial functional traits can be gained from 16S rRNA gene–based approaches by FAPROTAX, which was introduced for ocean environments and was recently improved and extended also for other environments (Sansupa et al. 2021).

In the past, bacterial communities were often considered as mere contaminants that can inhibit algae productivity or terminate algae populations. However, recent findings have shown positive effects of algae–bacterial interactions considered promising for biotechnological applications. The potential benefit can be found in enhanced biomass production, simplified harvesting, wastewater treatment, and bioremediation for sustainable aquaculture (Wang et al. 2014; Fuentes et al. 2016).

Apart from bacterial communities, algae also have to face fungus-like parasites, which pose a high threat for plant managers as they cause a sudden and massive death of microalgal cells in natural environments as well as in industrial systems (Gachon et al. 2010; Carney and Lane 2014). The mycoloop, driven by such zoosporic parasites, has a strong ecological impact on aquatic food web dynamics by regulating phytoplankton population size and providing access of nutrients to higher trophic levels (van Donk 1989; Sime-Ngando 2012; Kagami et al. 2014; Rasconi et al. 2014; Frenken et al. 2017; Jephcott et al. 2017). The family Aphelidiaceae represents a poorly known group of obligate endoparasitoids of various algae. The group was recently affiliated to the superphylum Opisthosporidia, but the phylogenetic relation to fungi is still under active discussion (Karpov et al. 2014; Tedersoo et al. 2018; Torruella et al. 2018). Hitherto, four genera (Aphelidium, Amoeboaphelidium, Pseudaphelidium, Paraphelidium) with 17 species have been described (Letcher and Powell 2019; Karpov et al. 2020; Seto et al. 2020). The distinct morphological feature in the complex lifecycle of aphelids has been determined as an intercellular amoeboid state. The detailed life cycle of Amoeboaphelidium protococcarum has been characterized by Letcher and colleagues (2013). Briefly, pseudopodiate aplanospores of A. protococcarum attach to intact host algal cell walls and form encapsulated cysts. Thereafter, penetration tubes are formed by the cysts to enter as ameboid trophont into the algal cytoplasm. The trophont engulfs the algal cytoplasm by phagocytosis, leaving a residual body. After division, a mature sporangium releases new flagellated zoospores of A. protococcarum.

The host specificity of A. protococcarum comprises protococcous algae, including various Scenedesmus species (Gromov and Mamkaeva 1968, 1969; Schnepf et al. 1971). Furthermore, different algae species or even strains exhibited varying sensitivity to different A. protococcarum strains (Gromov and Mamkaeva 1969). In 2017, a new isolate of A. protococcarum (WZ01) was identified causing an algal population collapse in an open raceway pond (Ding et al. 2017). Additionally to some unusual morphological characteristics, shifted pathogenicity patterns towards coccoid green algae were observed in the WZ01 isolate. These results underline that a comprehensive understanding of host specificity and the underlying mechanism of how aphelids detect and differentiate host cells is still not reported.

Infections with zoosporic parasites can lead to rapid and complete biomass loss in large-scale algal plants, which has been reported repeatedly in literature in recent years (Hoffman et al. 2008; Letcher et al. 2013; Carney et al. 2014; Ding et al. 2017). As commercial scale culture systems tend to be unialgal, specific parasites have a devastating effect on mass monocultures. Contamination management is still at an early stage, but promising methods of synthetic ecology and community engineering that manage microbial diversity and species composition leading to increased resilience and productivity simultaneously come to the fore (Shurin et al. 2013; Smith and Crews 2014; McBride et al. 2016). Hence, it is crucial to further investigate and understand the co-occurrence and microbial associations in artificial microalgae systems to accelerate such approaches (Mendes and Vermelho 2013; McBride et al. 2014; Carney et al. 2016).

Objectives of this study were (i) to investigate the temporal development of bacterial microbiomes within cultures of different algae species; (ii) to compare the effects of sterile laboratory conditions (sterile) and nonsterile culture conditions usually applied in industrial production processes (open) on the bacterial community composition; and (iii) to assess whether, or to what extent, associated bacterial communities are affected if the respective algae has been infected by aphelids (+ AI).

Material and methods

Algae/parasite culturing and experimental design

Algae cultivation was conducted in closed photobioreactors at the Competence Center Algal Biotechnology in Koethen, Germany, in 2017/2018. All experiments were performed in bubble column reactors (BCR) with 1.5L capacity and autoclaved modified Bolds basal medium (BBM) (Bischoff and Bold 1963) at 23.5 °C, a gas flow of 1.0 volume of air per culture medium (1% CO2) and permanent illumination at 100 μmol photons m−2 s−1 with a daylight white LED light (380–750 nm).

Experiments were conducted with the algae strains Scenedesmus vacuolatus SAG 211-8b, Desmodesmus quadricauda CCALA 464, Chlorella sorokiniana SAG 211-8 k, and Botryococcus braunii var. Showa. In each case, BCRs were inoculated to an optical density (OD) of 0.2 at a wavelength of 750 nm under sterile laboratory conditions (sterile; handling in a laminar flow box) and open “industrial” conditions (open; handling on a workbench), each with 6 replicates. Three independent replicates of each algae culture were thereafter infected with 100-mL (6.6% v/v) aphelid inoculum (+ AI), while 100-mL ddH20 was added to additional three independent replicates as negative control cultures (-AI). Each algae culture of the four resulting culture treatments (sterile + AI, open + AI, sterile -AI, open -AI) was sampled every 48 h for a period of 21 days and immediately checked for growth parameters. Dry weight content was determined by gravimetric measurement of dried biomass (Moheimani et al. 2013), and OD750 nm was detected on a Tecan microplate Reader (Tecan; Switzerland). Nutrient consumption was monitored by ion chromatography using a Dionex ICS 1100 system (Thermo Scientific; USA) according to manufacturers’ instructions. The morphological status of algal cells was evaluated and documented by using bright-field microscopy with an Olympus BX41 microscope and Olympus XC50 camera. Retained samples for nucleic acid–based analysis of the bacterial community structure were stored at − 80 °C until further processing.

An enrichment culture of A. protococcarum was obtained from a contaminated Christmas tree bioreactor at the Biosolar Center in Koethen, which was described previously (Matthes et al. 2015). Strain A. protococcarum AI15TR was isolated through repeated serial dilution and plaque plating on BBM agar plates as explained earlier (Letcher et al. 2013), and the elimination of other microorganisms was assessed by microscopic observations and cultivation with R2A selective agar as described earlier (Gibbs and Hayes 1988; Andersen 2015). The resulting aphelid stock cultures were stored at − 20 °C for longer periods without losing the infectivity to S. vacuolatus. To prepare aphelid inoculum, axenic cultures of S. vacuolatus were grown to mid-log phase in BBM for 5 days, diluted to a final OD750 of 0.2 in 1.4 L, infected with 100 mL purified A. protococcarum stock and cultivated in BCRs using the same culture parameters applied in the following experiments. Seven days after infection, cultures were harvested, microscopically checked for infection status and thereafter frozen for later use. For all contamination experiments described in this publication, the same bulk volume of aphelid inoculum was used to secure consistent composition, cell density, and pathogenicity of A. protococcarum AI15TR.

DNA extraction and amplicon sequencing

After 0 days, 8 days, and 16 days of incubation of each treatment, samples were collected, and independent replicates of the samples were pooled in equimolar amounts. One sample of the aphelid inoculum was analyzed alongside the algae samples. Genomic DNA was extracted according to Letcher et al. (2017) using the DNeasy Powersoil Kit (Qiagen; Netherlands) following the manufacturer’s protocols. Quantity and quality of nucleic acid extracts were checked and sent to Microsynth AG (Balgach, Switzerland) for subsequent amplicon sequencing. Two-step Nextera PCR libraries of the V5 to V7 regions of the bacterial 16S rRNA gene were employed by using the primer pair 799F (5′-AACMGGATTAGATACCCKG-3′) and 1139R (5′-ACGTCATCCCCACCTTCC-3′) (Beckers et al. 2016). Subsequently a v2 500 sequencing run on the Illumina MiSeq platform was performed according to the manufacturer’s instructions (Illumina; USA) by Microsynth AG (Balgach, Switzerland).

Sequencing data analysis

Bacterial 16S rRNA gene sequencing data were processed according to Buettner and Noll (2018) resulting in a final absolute OTU table. OTU sequences were compared against the reference sequences of the SILVA 16S database, and taxonomies were predicted using SINA aligner (v1.2.11), considering a minimum confidence threshold of 0.7 (Pruesse et al. 2012). Bacterial OTUs that could not be affiliated with a higher resolution are hereafter referred to as “unknown bacteria”. A MUSCLE alignment (Edgar 2004) was calculated based on the neighbor-joining method (Saitou and Nei 1987) in MEGA X (Kumar et al. 2018). The alignment was used to compute a phylogenetic tree with the maximum likelihood method using the general time-reversible model (Waddell and Steel 1997). Relative abundances were calculated on basis of the absolute OTU table and used for subsequent analyses.

Bacterial community composition

Composition and structure of bacterial communities based on sequencing data were analyzed using the bioconductor package phyloseq (v 1.28.0) and visualized with the implemented package ggplot2 (v3.3.2) on the open-source platform R (v3.6.1). Calculations of alpha diversity measurements were conducted using the Shannon index as an estimator of species richness and evenness, while Chao1 was used for estimating the number of species in a community based on the low abundance species (Kim et al. 2017). To determine the structure of the bacterial community composition, Bray–Curtis dissimilarity and weighted UniFrac distance metrics were computed and visualized by nonmetric multidimensional scaling (NMDS). While Bray–Curtis dissimilarity includes only abundance data, the weighted Unifrac metric also includes phylogenetic information to calculate distances (McMurdie and Holmes 2013). In addition, a PERMANOVA test (ADONIS) was conducted in the vegan package (Oksanen et al. 2019) to statistically test the effect of algae species, incubation time and culture treatment on the bacterial community structures. Relative OTU abundances of the bacterial communities on genus level were visualized by displaying the most abundant 20 genera. Rare genera, which accounted for < 1% each in each algae data set, were merged into the category “Accumulated < 1%”. Furthermore, numbers of shared and unique OTUs present in the bacterial community composition of different sample types (algae species or culture treatments) were determined by Venn diagrams (Larsson 2020).

Microbial community function analysis

Metabolic and ecologically relevant functions were retrieved from the FAPROTAX database (Louca et al. 2016) by using the MACADAM explore web interface at genus level (Le Boulch et al. 2019). Subsequently, functions from all bacterial genera were compiled per sample and assigned to different groups (algae species and culture treatments). Finally, the percentage of each function for each group was calculated with Excel (Microsoft, USA) and plotted with the ggplot2 package on R.

Results

Amoeboaphelidium protococcarum decreased algal growth parameters

Dry weight, optical density, and nutrient consumption of all noninfected algae cultures (-AI) continuously increased over time irrespectively of culture conditions (open vs. sterile) (Fig. 1, Table S1–3). On the other hand, an algal species–specific decrease in the same parameters was observed after aphelid infection (+ AI) (Fig. 1, Table S1-4). Morphological changes of algal cells like attached cysts, chloroplast reduction, and residual bodies as typically induced by intracellular parasitosis of A. protococcarum were observed in infected S. vacuolatus, D. quadricauda, and C. sorokiniana, but not in B. braunii (Fig. S1).

Fig. 1
figure 1

Analysis of dry weight biomass content of S. vacuolatus, D. quadricauda, C. sorokiniana, and B. braunii cultures over 21 days of incubation. Comparison of infected (+ AI) and noninfected cultures (-AI) in open and sterile conditions are shown respectively in the legend (mean of n = 3 ± SD)

Full size image

Bacterial community compositions were shaped by algal host and diverged after aphelid infection

Bacterial community structures were shaped by the respective algal host as determined by weighted UniFrac (Fig. 2, Table 1) and Bray–Curtis distances (Fig. S2, Table 1). Overall bacterial composition within one algae host was similar to each other at the beginning of the experiment but significantly diverged over time (Fig. 2). Aphelid infection had a greater impact on shifts of the bacterial community composition compared to culture conditions (Fig. 2). Both Shannon and Chao1 indices increased over time for all aphelid infected algae cultures (+ Al) regardless of culture conditions (open vs. sterile conditions) (Fig. S3). Bacterial diversity was lower in algae cultures without aphelid infection compared to those with aphelid infection (Fig. 3). Culture conditions had only a minor effect on the bacterial diversity associated with D. quadricauda and B. braunii, but bacterial diversity was more complex in open culture conditions when associated with S. vacuolatus and C. sorokiniana (Fig. S3), especially in noninfected cultures.

Fig. 2
figure 2

Nonmetric dimensional scaling (NMDS) based on weighted UniFrac distances of bacterial community composition retrieved from S. vacuolatus (purple), D. quadricauda (cyan), C. sorokiniana (green) or B. braunii (red). Algae cultures are marked according to culture treatments (circle ( ) means open-AI; triangle ( ) means open-AI; square ( ) means sterile-AI, cross ( +) means sterile + AI) and sampling times (days 0, 8, 16). Ordination stress indicates the goodness of fit

Full size image
Table 1 Effects of algal host, incubation time, and culture treatments on the bacterial community composition as revealed by PERMANOVA analysis of the Bray–Curtis dissimilarity and Unifrac distance metrics
Full size table
Fig. 3
figure 3

Succession of bacterial community compositions in four microalgae cultures at genus level. Rare genera, accounting for less than 1% relative abundance in each algae data set, were summarized in “Accumulated < 1%”. “Unknown Bacteria” represent OTUs that could not be assigned at phylum level. Relative abundance is displayed for four culture treatments (open-AI, open+AI, sterile-AI, sterile + AI) over time. ANPR, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium

Full size image

The overall bacterial composition was dominated by members of the phyla α-Proteobacteria, γ-Proteobacteria, and δ-Proteobacteria (for details see Table S5). The profound majority of taxa on OTU level only occurred in samples of one algae species (Fig. S4). In contrast, approx. 48.0% of bacterial OTUs were present in every algae culture regardless of the treatment. Additionally, 26.8% of the communities were only found in samples treated with aphelid inoculum (+ AI), while OTUs occurring only in samples of noninfected algae cultures (-AI) accounted for 43.5%.

Although shifts in the bacterial community composition were attributed to aphelid infection and incubation time, several dominant taxa remained stable for the respective algal host (Fig. 3). Likewise, bacterial community composition associated with S. vacuolatus was mostly composed of members of the genera Hydrogenophaga, Sphingomonas, and Brevundimonas (Table S5). However, the bacterial community composition of S. vacuolatus was less diverse in sterile and open culture conditions without inoculum addition (Fig. 3, Fig. S5). Members of the genus Brevundimonas and Devosia dominated the open and sterile conditions after 2 and 4 days, respectively, and thereafter, the community composition pattern was more diverse. Bacterial community composition associated with C. sorokiniana was characterized by members of the genera Porphyrobacter, Brevundimonas, and Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium (ANPR) group. In contrast, bacterial community composition retrieved from D. quadricauda was characterized by members of the genera Sphingopyxis, Blastomonas, and Brevundimonas, while B. braunii microbiomes were composed of Porphyrobacter, Sphingopyxis, and Hydrogenophaga (Table S5).

Since the growth of S. vacuolatus cultures was significantly less affected by the infection with A. protococcarum in open culture conditions, the major differences of open and sterile microbiomes were compared (Fig. 4). Especially relative abundance of Hydrogenophaga increased in open culture conditions whereas, e.g., Pseudomonas and Sphingomonas showed higher relative abundances in sterile culture conditions once A. protococcarum infected S. vacuolatus (Fig. 4).

Fig. 4
figure 4

Shift in the relative abundances of bacterial genera retrieved from infected S. vacuolatus cultures (+ AI) incubated under open and sterile culture conditions over time. Negative shift in relative abundances of respective bacterial genus (red) indicate increased abundance in the sterile culture conditions, while positive shift (green) indicate increased abundance in the open culture conditions. ANPR, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium

Full size image

Aphelids caused shifts in bacterial functional traits

The bacterial functional traits associated with all algal hosts included photoautotrophic and/or photoheterotrophic members (Table S6), which was understandable in a permanently illuminated, but nutrient-poor environment. In addition, all algal cultures were inhabited by chemoheterotrophic bacterial members. Furthermore, functional traits to metabolize mineral nutrients such as manganese, nitrate, sulfur, or iron components, which are key elements in algae culture media, were prominently represented. The overall functional diversity was in great accordance between the algae species (Table S6), despite the compositional differences of the bacterial community. However, if aphelid infection was present, the functional diversity of the bacterial community composition shifted and bacterial functions like detoxification, dark oxidation, oil bioremediation, cellulolysis, degradation, and chitinolysis were found more frequently in algae cultures with aphelid infection (Fig. 5).

Fig. 5
figure 5

Bacterial functional diversity in infected (+ AI) and noninfected (-AI) algae cultures. The presence of each predicted function was calculated for every sample belonging to both sample groups and deviations between groups over 10% were displayed as a bar plot and were based on FAPROTAX according to Louca et al. (2016). More information to the functions can be found in Supplementary Table S4

Full size image

Discussion

As industrial plant operators rarely publicly publish reasons or frequency of algal culture breakdowns, the financial impact of aphelid contaminations is difficult to assess. This study investigated the temporal bacterial community composition of four microalgal cultures under open and sterile culture conditions in presence and absence of aphelid parasites. Distinct bacterial communities for each algae species were observed, which only sparsely shifted over time in different cultivation conditions. Once aphelids were present, bacterial diversity increased, and algae rapidly lost their chlorophyll and biomass regardless if sterile laboratory conditions or nonsterile culture conditions were applied. Moreover, aphelid infection changed the bacterial functional profiles, which implicate a fundamental breakdown in the bacterial–algae interaction patterns.

Algae species–specific bacterial microbiota

Bacterial community compositions associated with algal hosts have previously been characterized by the dominance of Proteobacteria, Bacteroidetes, and Actinobacteria (Lakaniemi et al. 2012b; Krohn-Molt et al. 2013; Ramanan et al. 2015; Geng et al. 2016b; Fulbright et al. 2018), which is in line with our findings on phylum level. However, previous studies were inconclusive on species specificity of phycosphere bacteria on finer taxonomic levels (Sapp et al. 2007; Ramanan et al. 2015). In this study, bacterial microbiomes were significantly adapted to their algal hosts at genus level (Figs. 2 and 3). Although most bacterial genera occurred frequently in varying abundances in all algae species, we have found distinct and stable abundance patterns for each algal host species. These findings support the idea of a distinct core microbiome with a close relationship to its algal host, which was recently supported by findings for Chlorella saccharophila, Scenedesmus quadricauda, and Micrasterias crux-melitensis and its associated bacterial community members (Krohn-Molt et al. 2017). Such implications were also suggested for epibiotic microbiota–diatom community patterns, which were observed as long-term associations in controlled laboratory conditions (Crenn et al. 2018). Moreover, Ramanan et al. (2015) concluded that overarching bacterial functional traits are essential for the maintenance of biogeochemical networks in the nutrient-rich phycosphere, which also directly affects algae livelihood and bacterial–algal interaction patterns.

Effects of culture conditions on bacterial communities

Furthermore, it is still under discussion, whether associated microbiomes in artificial algae cultures originate from sampling habitats and might be preserved for years in strain culture collections or whether bacteria might be invading axenic strain cultures while nonsterile upscaling procedures were carried out.

Our findings clearly showed the presence of persistent algae-specific microbiomes from the beginning of the experiments, irrespective of open or sterile culture conditions. Precultures from Koethen’s strain collection have been handled with or without contact to the open environment in experimental preparation and upscaling, so immigration of new bacterial strains was very likely only in open culture conditions. However, open conditions had no impact on the algal growth and in most cases only a minor impact on the bacterial community composition (Figs. 1,2, and 3), indicating that invasion of nonalgal specific microbiome members is less successful. Such maintenance in the bacterial core microbiome indicates a high degree of adaption and robustness to the specific holobiont. However, since the bacterial community composition of S. vacuolatus exhibited the most tremendous changes over time, more incubation time points were analyzed to understand the succession of the bacterial compositions (Fig. S5). Members of the genus Brevundimonas, which dominated open cultures, were previously found to promote the growth of algal hosts (Park et al. 2008). Members of the genus Devosia and Blastomonas dominating the sterile cultures of S. vacuolatus were described with high transcription activity in the biofilm of a Scendesmus species linked to algae effector molecules (Krohn-Molt et al. 2017). These findings indicate that shifts in the algal culture conditions were associated with some bacterial genera that support algal growth in changing culture conditions. After acclimatization of the algal host, a broader spectrum in the matured bacterial community composition was established which supplement the abundant members of the core microbiome. In turn, the presence of aphelid parasites triggered a different and more diverse bacterial pattern of the algal host S. vacuolatus (Fig. 3, Fig. S5). The results of multiple long-term outdoor studies also suggested the persistence of a core microbiome but, in contrast to our experiments, observed seasonal changes (Biondi et al. 2017) or growth phase dependent changes of parts of the community (Geng et al. 2016b; Fulbright et al. 2018). Behringer and colleagues (2018) compared the composition of short-term cultivation of natural diatom isolates and long-term cultivation of strain collection cultures of diatoms. Results of this study suggest that in controlled laboratory culture conditions, interspecies interactions select for a specific, but simplified microbial community, shaped by long-term associations of diatoms and bacteria (Behringer et al. 2018). A recent study focused on the question of how bacterial community assembly processes are organized in the microalgal phycosphere of Phaeodactylum tricornutum and Microchloropsis salina in outdoor raceway ponds and laboratory enrichments (Kimbrel et al. 2019). Results showed that algal-associated bacterial communities were not only shaped by deterministic processes like algal host interactions and culture conditions, as reported in our study, but also by stochastic processes like initial inoculum composition of the algal microbiome.

Impact of parasitic infections on bacterial communities

The effects of zoosporic parasites in algae cultivation systems for industrial applications are, so far, poorly assessed even though these interactions were known from natural environments (Gerphagnon et al. 2013; Grossart et al. 2016; Frenken et al 2017). This study assessed how indigenous bacterial microbiomes of algae cultures were affected by an aphelid parasitosis. Not only algae growth was inhibited by the infection, but the bacterial community was also clearly afflicted in this process. Bacterial diversity increased (Fig. 3), and the community composition was altered over the course of aphelid infection in structure and functional profiles (Figs. 2, 3, and 5). Bacterial functions related to biomass breakdown were detected more often in infected algal cultures compared to noninfected ones (Fig. 5). This may result from an increasing number of dead algae biomass that was released by aphelids and became bioavailable.

There are some reports that bacterial community composition can be altered by antibiotic treatment and quorum sensing compounds (Geng et al. 2016a) and under the attack of ciliates (Lakaniemi et al. 2012a), but to our knowledge, there is only very few data published that specifically focuses on the effect of algae parasites on associated microbial communities in industrial cultivation systems. For instance, one study by Carney et al. (2014) monitored the dynamics of the microbiome in a microalgal mass culture growing in municipal wastewater. They reported an algal biomass decline by chytrid parasitosis related to a simultaneous shift in the bacterial community composition (from γ-Proteobacteria to Bacteroidetes) in a municipal wastewater photobioreactor, supporting our hypothesis that parasitic infections not only impact the algal population but also affect the associated microbial community.

Conclusions

Our study showed robust and temporal stable community patterns of the algae host and its associated microbiome were shifted once algal hosts were invaded by aphelids. Moreover, the functional traits of these microbiomes also shifted from algae-cooperative to decomposing traits. So far, algal target sites for initial aphelid invasion are unknown; therefore, more research is needed to understand how algae target sites can be protected and/or the aphelid actions can be disturbed or weakened by algae–bacterial interactions (McBride et al., 2016). Upcoming scientific research should address the microbial–algal interaction patterns during aphelid infection at the proteomic level in order to elucidate the cross-kingdom organismal link at the physiological level. Moreover, the advantage of temporal and spatial diversification of algal crops should be addressed, which may mitigate algal loss in case of an aphelid infection.