Introduction

Biofilms are cellular communities attached to surfaces and embedded in self-produced extracellular polymeric substances (EPS). In the broadest sense, all kinds of microbial structures like cell aggregates (flocs), sedimented layers or tightly attached films can be defined as biofilms [1, 2]. They are omnipresent in nature, where they are often found as multispecies microbial mats. In such structures, a complex system of interrelated dependencies of various organisms comprising phototrophic and heterotrophic strains exists. Biofilms are also very important in artificial medical or technical settings, although there they are mostly unwanted, as biofilms are very persistent and lead to biofouling or severe infections. However, these characteristics could be beneficial in terms of process stability and a continuous process mode for biocatalysis and the production of value added compounds with biologically challenging reactants as shown recently [3]. Higher overall robustness, self-regeneration and immobilization with high biomass retention are main characteristics for catalytically active biofilms [4, 5]. Key challenges when applying catalytic biofilms in continuous processes include mass transfer through the three-dimensional structures of cells and EPS substances and excessive biomass formation resulting in clogging. One approach to circumvent these bottlenecks is the formation of a flat, compact film. This can be achieved by cultivating the respective microbes in capillary tubings under continuous flow applying hydrodynamic forces. These may be generated by two-phase segmented flow, consisting of an aqueous and a gaseous phase [6]. The heterotrophic model strain used in this study, Pseudomonas taiwanensis VLB120, had to adapt over several days to the high interfacial forces, and formed a completely different structure under segmented as compared to single phase flow. This was highly beneficial for the respective biotransformation rates and provided a basis for a truly continuous process format. While applications of heterotrophic biofilms are well described, examples using photoautotrophic organisms as biofilm catalysts are missing up to date, although numerous studies reported proves of concept for the synthesis of value added compounds by suspended phototrophic organisms like cyanobacteria and/or microalgae [7–13].

In ecology, natural phototrophic-heterotrophic multispecies biofilm communities play a major role for ecosystem assessment and are frequently studied [14–19], whereas studies of phototrophic biofilms for synthetic and technical applications are limited [20]. Examples include microbial fuel cells [21] and applications in waste water treatment [22]. While in waste water treatment huge amounts of biomass are generated, microbial fuels cells represent a static non-flow system without hydrodynamic forces. Such systems are not suited for applications in whole cell biocatalysis due to mass transfer limitations and uncontrolled biomass growth [23]. The development of suitable cultivation concepts and the identification and characterization of monospecies biofilms will be essential for applications of phototrophic biofilms as catalysts in synthetic chemical industries.

We present growth controlled cultivation applying a segmented type of flow and the initial characterization of a monospecies Synechocystis sp. PCC 6803 biofilm in a micro capillary flow through reactor. Significant differences in biofilm growth could be observed depending on the hydrodynamic conditions applied. While under single phase flow, clogging of the capillary occurred due to uncontrolled biomass formation, a photosynthetically active, flat and compact biofilm structure was formed upon the introduction of air segment into the system without any cell detachment events.

In future combinations with metabolic engineering of Synechocystis sp. PCC 6803, this biofilm reaction setup may enable continuous light driven syntheses of value added compounds.

Materials and methods

Chemicals and bacterial strains

All chemicals used were obtained from AppliChem (Darmstadt, Germany), Sigma Aldrich (Steinheim, Germany), or CarlRoth (Karlsruhe, Germany) with the highest purity available. The cyanobacterial strains Synechocystis sp. PCC 6803 [24] and Synechococcus elongatus PCC 7942 [25] were obtained from the Pasteur Culture Collection of Cyanobacteria (PCC, Paris, France).

Cultivation of cyanobacterial strains

Precultures for biofilm inoculation were grown in 50 mL baffled shaking flasks in a Multitron shaker (Infors, Bottmingen, Switzerland), equipped with light tubes (Osram Lumilux L18 W/840 cool white) and humidity control. 10 mL of YBG11 medium [26] was inoculated from a cryo stock and strains were cultivated at 30 °C, 150 rpm, 80 % relative humidity and 50 µE m−2 s−1 photosynthetically active radiation (PAR). For long-term storage the strains were frozen in YBG11 medium containing 8 % DMSO at −80 °C.

Experimental biofilm setup

Biofilms were cultivated in rectangular-shaped borosilicate glass capillaries (Fig. 1) suitable for CLSM investigation [6]. YBG11 medium was supplied through Tygon Standard R-3607 tubing (2.06 mm inner diameter, Ismatec, Wertheim-Mondfeld, Germany) by a peristaltic pump (ISM 930; Ismatec, Wertheim-Mondfeld, Germany). In case of segmented flow, additional tubing was connected via a T-connector and a sterile filter (0.2 µm) for the introduction of air segments. The assembled flow cell was heat sterilized. Afterwards the system was flushed with YBG11 and adjusted to 30 °C and 55 µE m−2 s−1. If not stated otherwise, the biofilm flow cells were inoculated with cells in 800 µL of preculture with an OD750 of 1.0. Cell attachment occurred in 24 h without flow. Subsequently, the medium flow was started. Two different flow patterns and their influence on biofilm development have been evaluated—single phase flow: After the cell attachment phase the flow was set to 100 µL min−1 (1.4 cm min−1) unless indicated otherwise. The flow was kept constant during the whole cultivation time; segmented flow: After the cell attachment phase the flow was set to 100 µL min−1 (1.4 cm min−1) and kept constant for 2 days. Subsequently segmented flow was established and air segments were introduced into the aqueous phase at 50 µL min−1 while the aqueous flow was reduced to 50 µL min−1, resulting again in an overall flow rate of 100 µL min−1 (1.4 cm min−1). For more details, please see text. An overview showing the reactor system is presented in Fig. 1.

Fig. 1
figure 1

Schematic representation of the biofilm reactor (capillary length 80 mm) operated in single phase flow (a) and segmented flow mode (b). Light is absorbed and scattered on its way through the individual biofilm layers, whereas CO2 is consumed by the cells resulting in a CO2 gradient. O2 produced by photosynthesis is leaving the system via the bulk phase

Full size image

Analytical procedures

Planktonic cell growth was quantified by turbidity measurements (OD750, Libra S11, Biochrom Ltd, Cambridge, UK). Photosynthetically active radiation (PAR) was determined with a MQS-B mini quantum sensor coupled to an ULM-500 light meter (Heinz Walz GmbH, Effeltrich, Germany).

Biofilm development on the glass surface was monitored in situ using a Zeiss LSM 5 Pascal confocal laser-scanning microscope (CLSM; Carl Zeiss, Jena, Germany), equipped with an argon and helium–neon laser and an EC Plan-Neofluar 20x, 0.50 Ph2 M27 objective, taking advantage of the natural autofluorescence of chlorophyll and the phycobilins phycocyanin, allophycocyanin and allophycocyanin-B [27, 28]. For photosynthetically non-active cells, the autofluorescence of carotenoids was detected as described earlier [28, 29]. Three-dimensional (3D) biofilm image reconstructions were performed by IMARIS (Bitplane AG, Zürich, Switzerland).

Results and discussion

Robust biofilm formation with Synechocystis sp. PCC 6803

For the investigation of biofilm formation, flow cells were inoculated with Synechococcus elongatus PCC 7942 and Synechocystis sp. PCC 6803, respectively. Under the conditions applied, Synechococcus elongatus PCC 7942 did not attach to the glass substratum. A reason why cell attachment and biofilm formation with this strain was not possible may be a mechanism for self-suppression of biofilm formation induced by a self-produced extracellular compound, as reported recently [30]. In contrast to that, initial cell attachment of Synechocystis sp. PCC 6803 to the glass surface of the flow cell reactor took place within 4 h. Since Synechocystis sp. PCC 6803 showed strong biofilm formation, biofilm growth and morphology were characterized under different hydrodynamic conditions.

Uncontrolled heterogeneous single species biofilm formation under single phase flow

No-flow, batch incubations in the tube allowed sufficient initial cell attachment overnight. Then, at an applied flow rate of 50 µL min−1 (0.7 cm min−1), exponential growth was observed during the first 72 h of the experiment at a growth rate of 0.045 h−1, calculated based on the increase in overall biofilm volume. A gradient in biofilm thickness toward a thinner film at the outflow was observed over the reactor length. Most probably this was due to carbon dioxide depletion at this low flow rate. Synechocystis sp. PCC 6803 formed a very heterogeneous biofilm under these conditions without an area-wide sublayer and with a porous structure interconnected by channels and elongated streamers. Overall, this resulted in a high surface ratio of 5.5 µm2 µm−2, and an inhomogeneous thickness and light distribution (Fig. 2a). After 7 days the flow rate was increased to 100 µL min−1 (1.4 cm min−1), preventing CO2 limitation. To exclude an influence of the carbon dioxide limitation in the beginning of the experiment on the overall development of the biofilm structure, we repeated the experiment with an initial flux of 100 µL min−1 (1.4 cm min−1). The same heterogeneous biofilm structure described before was observed.

After 14 days of constant flow, the capillary started to clog and as a result the system pressure increased and a constant flow was not possible anymore. At this time point the biofilm was in large parts still permeable to light, but could not be supplied with CO2 any longer.

Fig. 2
figure 2

Representative images showing the differences in development of Synechocystis sp. PCC 6803 biofilms grown under single phase flow (a) and segmented flow (b). The first picture in every row shows a photo of the fully grown biofilm inside the glass capillary reactor. The following illustrations are 3D-reconstructions of the biofilm generated from CLSM data by IMARIS. The top views on the biofilm surface as well as cross sections of the biofilm are included. The first pictures were taken directly after starting the flow. In case of segmented flow the biofilm thickness at day 6 and 14 exceeds already the penetration capacity of the applied LSM technique

Full size image

Segmented flow shapes and stabilizes single species biofilm development

The capillary plug flow reactor was inoculated with Synechocystis sp. PCC 6803. After initial overnight cell attachment under batch conditions (non-continuous) and 2 days of single phase flow (100 µL min−1, 1.4 cm min−1), air segments were introduced as described in the method section. The air segments generated interfacial forces between 1.36 × 10−7 and 1.36 × 10−6 N (calculated according to [6]), in addition to a better availability of the substrate CO2. Surprisingly, this did not result in a significant detachment processes as observed for Pseudomonas under segmented flow [6]. In fact, no wash out of cells could be detected throughout the experiment. The attachment forces of the cells to the substratum are possibly much higher for Synechocystis sp. PCC 6803 compared to the heterotrophic Pseudomonas strain reported earlier. There, the interfacial forces generated by the segmented type of flow were strong enough to nearly completely detach all cells from their growth surface [6]. For Synechocystis a growth rate of 0.045 h−1 was determined during the first 72 h, which is comparable to the growth rates without segmented flow described above. However, segmented flow changed the biofilm structure dramatically. If the CO2 availability had any impact regarding the change in biofilm structure remains to be elucidated. There are no indications for a CO2 limitation in either system [single (100 µL min−1)/two-phase flow], as for both equal biofilm thicknesses from the beginning to the end of the plug flow reactor were determined, pointing toward a sufficient supply of this substrate.

During the first 14 days of the experiment, a homogenous, dense biofilm layer developed (Fig. 2b). The autofluorescence intensity, used for quantification of biomass development, is significantly higher compared to the biofilm grown under single phase flow. The signal intensity determined from the biofilm was depending on cell density and pigment concentration within the cells. Especially the latter is subject to changes due to adaptation to different light conditions and is influenced by laser irradiation of the CLSM. Nevertheless, this signal may be used for a qualitative assessment of the biofilm biomass.

At some point of growth, the biofilm thickness and the density of the mature biofilm grown under segmented flow exceeded the penetration capacity of the CLSM system and the biofilm thickness could no longer be determined (Fig. 2b). Yet, it was possible to characterize the further development of the biofilm by eye. Within 3 weeks, the biofilm grown in the capillary appeared nearly black. In contrast to the non-segmented aqueous microreactor described above, the biofilm reactor could be continuously operated for 5 weeks before it was actively terminated. Neither was there an indication that the biomass increased further during this time period, nor were any cells detectable in the flow through. Phototrophic segmented flow capillary biofilm reactors are therefore powerful tools to control and even prevent excess biomass formation and to keep the process controllable.

Key parameters affecting Synechocystis sp. PCC 6803 biofilm development

Key parameters for biocatalytic and biotechnological applications of cyanobacterial biofilms are the distribution of light and the diffusional limitation of carbon dioxide within the biofilm. In case of heterotrophic biofilms, the efficient supply of nutrients is a matter of diffusion, in most cases directly from the aqueous bulk phase. In contrast, light availability, essential for the growth of phototrophic biofilms, is determined via light scattering within the biofilm.

In natural phototrophic-heterotrophic mixed species biofilm communities, the special distribution of photo- and heterotrophic organisms, the attenuation of light intensity and the photosynthesis rate within these microbial mats have been intensively investigated [14–16, 31]. Depending on the densities of the investigated mats specific light attenuation coefficients were reported, which gave a basis for estimating respective values for the here cultivated Synechocystis biofilms (Fig. 3). Related to the observed structure of the biofilm growing in the microcapillaries, light attenuation coefficients between 7 and 12 mm−1 were estimated.

Fig. 3
figure 3

a Exponential light attenuation within a cyanobacterial biofilm calculated from data published for an artificial cyanobacterial mat [17]. The gray areas represent the estimated light attenuation for incidental light intensities of 55 and 150 µE m−2 s−1 with a light attenuation coefficient between 7 and 12 mm−1 (right and left borders, respectively). Additionally the results for an attenuation coefficient of 9 mm−1 are highlighted (- - -). Horizontal line represents the boundary of 15 µE m−2 s−1 where growth of Synechocystis sp. PCC 6803 is significantly decreased. b, c Schematic distribution of the main substrates carbon dioxide and photosynthetically active radiation (PAR) within the biofilm and the resulting potential growth zones

Full size image

The light attenuation coefficient together with the intensity of the available incident light defines the borders of growth in three-dimensional structures as biofilms. For the cyanobacterium Oscillatoria agardhii these limits have been closely evaluated. Below 20 µE m−2 s−1 a significant decrease in the growth rate was observed, while above 150 µE m−2 s−1 light inhibition occurred [32]. Based on these values a maximal thickness for Synechocystis biofilms cultivated under the herein described conditions has been calculated. Assuming the maximal possible light intensity of 150 µE m−2 s−1 and a light attenuation coefficient of 9 mm−1, which can be assumed for a biofilm grown under single phase flow conditions, a maximal biofilm thickness of 250 µm can be theoretically achieved. The maximal biofilm thickness to be expected under our experimental conditions (55 µE m−2 s−1) during single phase flow was about 150 µm (Fig. 3a). The light attenuation coefficient under segmented flow is significantly higher because of higher cell densities. With a theoretical light attenuation coefficient of 12 mm−1, the expected biofilm thickness would be around 105 µm, which is far beyond the detection limit of the available CLSM system. Therefore, after 5 weeks of segmented flow biofilm development, cross sections have been prepared from the capillary and were investigated by CLSM. This allowed the determination of a final maximal biofilm thickness between 70 and 120 µm. Biofilms are heterogenic, dynamic, three-dimensional structures comprising areas with high cell densities, low cell densities, and channels without any cells. These differences in the composition strongly influence the light attenuation in the biofilm and thus the possible biofilm thickness. Interestingly, the biofilm cross sections were made up nearly exclusively of photosynthetically active cells and only a few cells (<1 %) without photopigments could be localized, homogenously distributed throughout the biofilm. The anticipated limit for light availability lies within the measured range. Also important, the observed differences in biofilm thickness are corresponding to the expected variances in light attenuation within the biofilm. The negligible biomass concentration in the outlet stream, in addition to the fact, that there were nearly no photosynthetically inactive cells detected at the inner-tube biofilm surface led to the conclusion that there is no significant biomass formation within the reactor anymore. Theoretically, one would expect biofilm growth up to a thickness, where one of its main substrates, carbon dioxide or light, becomes limiting. Thus, maximal light and carbon dioxide penetration depth would define the borders of biofilm growth (Fig. 3b). This opens the question why further growth was not observed in our experiments. One explanation might be a quorum sensing signal, which prevents growth within the center of the biofilm, in the presence of high cell density [33]. Additionally, the oxygen produced within the biofilm might also be a reason for limited growth. A high oxygen concentration would have a negative impact on the carbon capture ability of the cells [34].

Reactor and process engineering for phototrophic whole cell biocatalysis require a rethinking of common reactor concepts. The essential substrates light, water, and carbon dioxide necessitate different process concepts and technical solutions compared to today´s applications in biotechnology. The available reactor configurations aim at maximal biomass production; however, this is usually suboptimal. This is also true for the suitability of biofilms as catalysts for continuous chemical synthesis which necessitates controllable porosity and growth as well as activity of all cells in the biofilms. For phototrophic microorganisms these qualities are for the first time met for the concept described herein. In our opinion, continuous processes, controlled biomass formation and regeneration of the biocatalyst are mandatory to enable the efficient production of value added compounds and biofuels from sunlight and CO2.

Concluding remarks

Biofilms are important functional modules in many synthetic bioprocesses. Their productivity is determined by stability, morphology, and control of biofilm development and growth. This report communicates a new methodology for developing, cultivating, and controlling a monospecies cyanobacterial biofilm of Synechocystis sp. PCC 6803 using a microcapillary reactor with continuous segmented flow. This concept provides a unique possibility to control biofilm growth and is a potent technology for the application of phototrophic cyanobacterial biofilms. This provides a basis for future investigations of biofilm growth and production characteristics also of other cyanobacterial strains. Open questions include scalability, molecular regulation influencing biofilm formation and solvent tolerance.