1. Introduction
Phycocyanobilin (PCB) is a bluish pigment covalently bounded to α and β subunits of some phycobiliproteins (PBPs) such as phycocyanin (PC) or allophycocyanin (APC) [
1,
2,
3]. Structurally, PCB is an open-chain tetrapyrrole that acts as these PBPs’ primary chromophore [
4]. Its spectroscopic properties enable specific taxa of photosynthetic organisms to utilize wavelengths near the red that are not useful to traditional pigments such as chlorophylls or carotenoids [
5]. Regarding their biosynthesis, PCB and other phycobilins (PBs), such as phycoerithrobilin (PEB) and phycoviolobilin (PVB), are derived from an alternative pathway of heme degradation. This pathway consists of two enzymatic steps, which, in the case of PCB, are catalyzed by two different enzymes [
6]. Firstly, heme groups are reduced to biliverdin IXα in the presence of O
2 by the heme oxygenase 1 (HO1). As a result of this catalysis, CO and Fe
+2 molecules are released (
Figure 1a). This first step is common to all pathways derived from the cleavage of the heme ring, encompassing both those related to the formation of all PBs and the purely degradative pathway that proceeds with the production of bilirubin and derivatives (urobilinogen, urobilin, stercobilin, etc.). Secondly, a phycocyanobilin:ferredoxin oxidoreductase (PCYA), reduces BV to PCB, passing through 18
1,18
2-dihydrobiliverdin (DHBV) as an intermediary compound (
Figure 1b). Both HO1 and PCYA are ferredoxin-dependent enzymes, and the electrons needed in each step are supplied by a ferredoxin-NADP
+ reductase (FNR), so the biosynthesis is strongly dependent on the reducing power availability (NADPH) in the cell.
In recent years, PCB has progressively increased its biotechnological relevance because multiple applications in the food industry have been proposed due to its natural blue color [
7]. Furthermore, numerous studies postulate that this pigment can become a powerful antioxidant and antitumor agent, which has generated significant commercial interest from the pharmaceutical and nutritional supplement industries [
3]. Some studies have even proposed PCB as a potential protector against Alzheimer’s disease due to its antioxidant, anti-inflammatory, and immunomodulatory properties [
8]. However, PCB production methods do not go beyond simple laboratory-scale research assays. Most of these methods try to break up the chromophore from the holo-PC through acidic hydrolysis and mainly through solvents such as methanol and ethanol, together with high temperatures [
9]. However, these methodologies are usually tedious, inefficient, and expensive, using solvents that can be dangerous for health and the environment. A second option for the industrial production of this PCB involves the heterologous co-expression of HO1 and PCYA in a standardized host such as
Escherichia coli BL21 (DE3) and coupling in vivo the enzymatic pathway using the host’s precursors. Recent works show several significant advances in this field, achieving variable production titers of PCB [
10,
11,
12,
13,
14,
15,
16]. Nevertheless, several key factors for the heterologous synthesis remain unclear, and frequently, the obtained results are not reproducible in other laboratories. In this study, the key factors (medium composition, O
2 availability, presence of precursors, concentration of the inducer, etc.) that determine the ability of the host to initiate, maintain, and preserve the production of this xenobiotic molecule are experimentally explored. Beyond this, we apply the main conclusions derived from 250 mL shake flask studies to test the PCB biosynthesis at the pre-industrial level using a 2 L bioreactor.
2. Materials and Methods
2.1. Strains and Genes
E. coli DH5α (ThermoFisher Scientific, Waltham, MA, USA) was used for molecular biology procedures, while
E. coli Bl21 (DE3) (Novagen-Merck, Darmstadt, Germany) was used as a host for the heterologous expression of proteins and as an enzymatic engineering scaffold. The genes encoding HO1 and PCYA enzymes were amplified from the
Synechococcus elongatus PCC7942 and
Synechocystis sp. PCC6803 cyanobacterium genomes. Both species were obtained from the ATCC culture collection (Manassas, VA, USA). After the first set of expression assays, the expression levels using native sequences were insufficient, and no color was detected in the pellets. For this reason, the codon usage was optimized for translation in
E. coli BL21 (DE3) using the GenSmart
TM online tool from GenScript Biotech Corporation (Pennington, NJ, USA) (
Figure S1). The resulting ORFs, named
ho11/
pcyA1 (from
S. elongatus) and
ho12/
pcyA2 (from
Synechocystis sp.), were artificially synthesized by the company Isogen Lifesciences B.V. (Veldzigt, The Nederthlands), including appropriate restriction sites for further manipulation.
2.2. Vectors and Genetic Constructions
The two ho1 genes were individually inserted in the MCS (Multiple cloning site) II of pCDFDuetTM vectors purchased from Novagen-Merck (Darmstadt, Germany). In the same way, two pcyA genes were harbored in the MCS II of pETDuetTM vectors from the same provider. Thus, all the possible combinations between the four different constructions could be assessed. Finally, a pcyA2/ho12 tandem was harbored in the MCS I and MCS II of the pETDuetTM vector. After confirming that the PCB titer was identical to that obtained with combined vectors, pETDuet–pcyA2/ho12 was chosen to carry out the subsequent assays.
2.3. Growth Media, Precursors, and Culture Systems
The components of the different growth media are described in
Table 1. All the reactives were purchased from AppliChem Panreac (Castellar del Vallés, Spain). Briefly, the assayed media used to improve the PCB synthesis were LB (Luria Bertani), TB (Terrific Broth), and MM9 (Modified M9 medium). MM9 medium contained 30 g/L of glycerol as the primary carbon source and 10 g/L of casein peptone as the nitrogen source. All the potential precursors were acquired from SigmaAldrich-Merk (Darmstadt, Germany). The assayed concentrations were 15 mM α-aminolevulinic acid (ALA), 10 µM Hemin (HEM), 15 mM glutamic acid (GA), 2 mM nicotine adenine dinucleotide phosphate (NADP
+), and 1 g/L of ascorbic acid (VC). To ensure the stability of the plasmids, the strains transformed with the pETDuet
TM and/or pCDFDuet
TM vectors were selected with 60 µg/mL ampicillin and 40 µg/mL streptomycin (AppliChem Panreac, Castellar del Vallés, Spain), respectively, or by a combination of them. Two different culture systems were employed in this study. The OD
600 of the cultures was determined using a 6131 Biophotometer (Eppendorf, Hamburg, Germany). On the one hand, the optimization of the PCB biosynthesis was conducted in 250 mL shake flasks containing 75 mL of medium. Conversely, all the batch assays in the bioreactor were carried out in a Biostat
® B (B. Braun Biotech, Melsungen, Germany) harboring a 2 L vessel with 1.2 L of MM9 medium.
2.4. Protein Expression and Biosynthesis Assays of PBs in Flasks and Batch Bioreactor
All synthesis assays started from a fresh pre-culture obtained by inoculating 75 mL of LB with a single colony from a fresh plate of transformants. As discussed later, temperatures close to 37 °C (optimal for the host’s growth) were detrimental to pigment synthesis. To avoid this issue, this inoculum was grown at 34 °C, which was the highest temperature in the subsequent optimization assays. This pre-culture was grown overnight at 180 rpm and used to inoculate the different expression assays at an OD600 of 0.1. After that, the cultures were grown under assayed conditions in each case until reaching the selected OD600 for the induction. As main determining factors for the PCB and BV synthesis, the effect of the induction OD600 (0.15, 0.3, 0.5, 0.8, and 1.0); the concentration of IPTG (isopropyl β-d-1-thiogalactopyranoside) as inducer (0.05, 0.1, 0.3, and 0.8 mM); the temperature of the culture (18, 28 and 34 °C) during biosynthesis, and the shaking speed (60, 180, and 250 rpm) were analyzed. When one of these four factors was examined, the other remained at their optimal points (28 °C, 0.1 mM IPTG, 250 rpm of shaking, and OD600 of 0.5). A control culture was kept under the same conditions for each assay, but the induction step was avoided. For the 2 L assays in a batch bioreactor, the culture was initially seeded at an OD600 of 0.25 and allowed to grow at 34 °C with 3 L/min of airflow and 200 rpm of stirring until reaching an OD600 of 1.5 units. At this point, it was induced with 0.1 mM of IPTG, and the temperature was reduced to 30 °C. After 4 h post-induction, the temperature was decreased to 28 °C to gradually slow the host’s metabolism. Due to the phosphate buffer incorporated into the MM9 medium, the pH remained constant between 6.8 and 7.2 units, so no corrections were needed throughout the assays. The pH, rpm, pO2, and temperature were continuously monitored during the batch. In both 250 mL flakes and 2 L batches, the total induction time was 8 h.
2.5. Monitoring of Pigment Synthesis
In all the assays, 2.5 mL of culture was taken every hour after induction. From this volume, 0.5 mL was employed to measure the OD600, and the remaining 2 mL were centrifuged. To avoid their degradation, the resulting pellets were stored at 4 °C until the visual analysis was carried out at the end of each assay. In this analysis, all collected pellets were visually compared with the control to establish a chronological synthesis line.
2.6. Quantification of PBs at Final Point
After 8 h of induction, the last OD
600 value was logged, and 50 mL of the culture was centrifuged. The pellet was cooled to 4 °C, and the pigment content was extracted by adding 3 mL of methanol, vigorous shaking, and incubation at 4 °C overnight. The methanolic extract was separated from the cellular debris by centrifugation (10 min, 14,000 rpm) at 4 °C. Finally, this extract was acidified by adding 5% (v/v) 10 N HCl. The samples were filtered to 0.2 µm and then analyzed by spectrophotometry or HPLC. For quantification, the extinction coefficient of 37.9 mM
−1 ·cm
−1 (ε
680) was used for PCB, and 30.8 mM
−1 ·cm
−1 (ε
695) was used for BV-IXα [
17]. For HPLC quantification, a Kinetex C8 column (Phenomenex, Torrance, USA) was employed under the control of an Agilent 1260 Infinity I system. The column was equilibrated with 50:50 water/acetone mobile phase and 0.06% acetic acid. During separation, 50 µL of acidified sample was isocratically eluted at 0.8 mL/min at room temperature for 18 min with the same solvent. The retention time for PCB and BV was 12.1 and 14.6 min, respectively, detected by a VWD (G1314B) detector at 600 nm. SigmaAldrich-Merck (Darmstadt, Germany) provided the PCB and BV standards to enable pigment identification and build the equipment calibration.
4. Discussion
In recent years, there has been a surge in the heterologous synthesis of bilins and PBs. Several studies have successfully synthesized both BV [
20,
21] and PEB [
22,
23], and PCB [
10,
11,
12,
13,
14,
15,
16] in
E. coli BL21 (DE3). Regarding the yield, the results reported in these works are highly variable, with production levels ranging from 3.0 to 184.2 mg/L [
10,
16]. In some of these studies, the genome of the BL21 (DE3) strain has even been modified to enhance the metabolic pathways for heme production as the primary precursor of these pigments and repress the genes involved in some of its degradation pathways [
12,
13,
16]. In other studies, the wild-type
E. coli Nissle 1917 strain has been used positively as an expression host following genetic manipulation. This strain can incorporate extracellular heme through a specific receptor/transporter (ChuA), thereby increasing the availability of the precursor [
21,
24]. In addition to the yield variability, these works show surprisingly variable results regarding the optimal parameters described for the production process. The present study offers valuable insights into the dependencies among key parameters influencing PCB synthesis rate and their impact on the host’s metabolism. On one hand, the metabolic state of the host seems crucial for achieving sustained pigment synthesis. In this sense, temperatures between 25 and 30 °C, combined with proper access to O
2 and abundant and assimilable carbon sources such as glycerol, resulted in an extended exponential phase and a considerable improvement in the rate synthesis and the stability of the pigment. Under these conditions, the host can maintain a high percentage of its electron carriers in a reduced state (NADPH), which are directly used to reduce exhausted FNR and continue supplying electrons to HO1 and PCYA. Simultaneously, this reducing power can keep the heme synthesis pathway active, where some enzymatic steps are also NADPH-dependent [
25].
In contrast, the approach to the stationary phase or any other source of stress halts the synthesis and, if prolonged, triggers a fast degradation process of the freshly synthesized pigment. In the obtained results, synthesis times longer than 8–10 h at 28 °C progressively returned the cells to the control phenotype, regaining their initial whitish appearance. BV and PCYA are exogenous molecules to the metabolism of
E. coli BL21 (DE3), so the metabolic pathways used by the host to degrade them and the reason for their activation when approaching the stationary phase are still unknown [
26]. On the other hand, the effect of the inducer on PCB biosynthesis has been highlighted. While IPTG concentrations below 0.1 mM resulted in excessively slow synthesis of HO1 and PCYA, concentrations above 0.3 mM gravely increased the host’s metabolic burden. This was likely due to the overwhelming focus of the BL21 (DE3) strain on the mass production of recombinant proteins, which ultimately damaged the cells by precipitating as inclusion bodies.
In summary, the availability of O
2, temperature, and the concentration of the inducer definitively influence the intensity of pigment production and the temporal duration of the biosynthetic process until the maximum concentration is reached. Thirdly, adding possible biosynthetic precursors for BV and PCB synthesis produced many results. Similar to other works, the presence of ALA as a heme precursor only generated slight differences in PCB production compared to the reference culture [
10]. Other potentially valuable precursors such as NADP
+ or VC did not generate differences, contrary to what is suggested in other studies in which both ALA and VC did improve the synthesis capacity, but not nicotinic acid [
11]. The addition of 10 μM of HEM caused a marked inhibition of bacterial growth and PCB synthesis, indicating a certain degree of toxicity and metabolic burden for the host. This result diverges from that obtained with
E. coli Nissle 1917, where the host could assimilate and incorporate extracellular heme as a precursor [
24]. Based on our results, among all the precursors tested, only GA, and to a lesser degree ALA, yielded significant improvements, not in the synthesis rate itself, but in maintaining it and preventing the degradation of the freshly synthesized pigment [
27].
Further research will determine why adding this amino acid generates a notably positive effect while ALA, a subsequent metabolite in the heme synthesis pathway, partially loses this effect. Lastly, it is necessary to highlight the potential connection detected in our results between the increased stirring speed (mediated by paddles) and the release of the fresh pigment into the culture medium. Several studies have previously reported the accumulation of both BV and PEB in the culture medium, however none of them offer an explanation for the cause of this release [
20,
23]. This is the first report that suggests a possible dependency between both parameters. The results reflect that O
2 availability significantly improves the synthesis rate of PCB. However, when O
2 was supplied by increasing the paddle speed, it caused mechanical stress, and the cells tended to excrete the pigment into the medium, probably due to membrane weakening and the subsequent cell death over the induction hours.
In summary, this study describes optimizing the heterologous synthesis of PCB in E. coli BL21 (DE3), achieving a maximum value of 3.8 mg/L. This value refers only to the pigment in the host biomass. However, it has been confirmed that the pigment location can be altered under specific situations (e.g., high paddle stirring), resulting in a significant concentration of excreted PCB into the medium or form foams. Consequently, this study contributes to the current understanding of the molecular mechanisms that control the synthesis, degradation, and heterologous mobilization of PBs in this host. These tests will facilitate the industrial production of these exciting molecules, making their application and commercialization possible soon.
5. Conclusions and Future Perspectives
In the reported work, we successfully coupled the enzymatic pathway and optimized the parameters for PCB synthesis in the standard host, E. coli BL21 (DE3). The best production results were achieved using a modified medium rich in glycerol and glutamic acid (MM9) in a 2 L batch bioreactor, oxygenated at 6 L/min with constant agitation at 200 rpm. The process was maintained at temperatures between 28 and 30 °C to maximize synthesis, and induction was performed by adding 0.1 mM IPTG when the OD600 reached 1.5 units. Under these conditions, a maximum PCB concentration of 3.8 mg/L was obtained 8 h post-induction. Although this production value is modest compared to yields reported in other studies, the parallel results of this study open new avenues for improving process efficiency. Despite suggestions from other studies that improving precursor availability is the key to increasing process yield, our work has demonstrated a strong dependence—equal to or more definitive than precursor availability—on dissolved O2 and the host’s ability to sustain PCB synthesis over time and avoid degradative processes associated with entry into the stationary phase.
On the other hand, this bioreactor model can maintain the dissolved O2 at a set point by raising the impeller speed while maintaining a constant airflow rate. However, while the amount of available O2 is maintained, this rise in mechanical stress leads to a disruption in PCB accumulation, causing host cells to excrete, either voluntarily or involuntarily, the pigment rather than store it. Further research is required to enhance O2 availability for the host without inducing mechanical stress. Designing an alternative sparging system for the bioreactor would enable the isolation of both variables (rpm and pO2) and assess how much the process yield can be improved by minimizing oxygen limitation. Such research is essential for the future large-scale industrial production of these pigments in a heterologous system.