Photoautotrophic production of renewable ethylene by engineered cyanobacteria: Steering the cell metabolism towards biotechnological use

Ethylene is a volatile hydrocarbon with a massive global market in the plastic industry. The ethylene now used for commercial applications is produced exclusively from non-renewable petroleum sources, while competitive biotechnological production systems do not yet exist. This review focuses on the currently developed photoautotrophic bioproduction strategies that enable direct solar-driven conversion of CO 2 into ethylene, based on the use of genetically engineered photosynthetic cyanobacteria expressing heterologous ethylene forming enzyme (EFE) from Pseudomonas syringae . The emphasis is on the different engineering strategies to express EFE and to direct the cellular carbon flux towards the primary metabolite 2-oxoglutarate, highlighting associated metabolic constraints, and technical considerations on cultivation strategies and conditional parameters. While the research field has progressed towards more robust strains with better production profiles, and deeper understanding of the associated metabolic limitations, it is clear that there is room for significant improvement to reach industrial relevance. At the same time, existing information and the development of synthetic biology tools for engineering cyanobacteria open new possibilities for improving the prospects for the sustainable production of renewable ethylene.


| COMMERCIAL SIGNIFICANCE AND PRODUCTION OF ETHYLENE: NEED FOR RENEWABLE INDUSTRIAL SOLUTIONS
Fossil resources continue to be extensively utilized for the manufacture of hydrocarbon fuels and different organic materials for human consumption, despite the severe environmental impact on the climate and global ecosystems. Due to the sheer volume of the use, which currently corresponds to over 4000 Mt of crude oil per year (IEA 2020), the detachment from petroleum-based products presents one of the grand challenges of humankind. Renewable energy sources such as wind and solar power now provide competitive alternatives for the fuel and energy sectors through the electrification of energy supply and transportation. However, we also need to find sustainable replacements for petrochemicals and derived materials, including solvents, plastics and synthetic fabrics, which constitute over 10% of the total crude oil demand (IEA 2018; for more on the subject see Levi & Cullen 2018).
The most abundant industrial petrochemical is the light olefin ethylene (C 2 H 4 ), which is globally used in chemical manufacturing at multimillion ton annual scale (Fernandez et al. 2018). Ethylene is primarily used as a precursor for producing polyethylene (PE), which is the most common plastic in the market, found for example in plastic bottles and other forms of packaging. Current largescale ethylene production relies on steam cracking of petroleum-derived materials, which requires high energy input and is one of the largest single CO 2 -emitting processes in chemical industry.
Although advanced reactor technologies are being developed to reduce the energy intensity and carbon emissions (Amghizar et al. 2020), the process is still based on oil, while alternative sustainable biotechnological production systems do not exist. This scenario does not only apply to ethylene but describes the global challenge in general. There is an urgent demand for entirely new carbon-neutral strategies to supply renewable precursors to the manufacturing industry, in parallel to the development of new products such as biodegradable plastics to replace those now in use (Pellis et al. 2021). Ultimately, this calls for complementary industrial solutions based on readily available sustainable sources of energy, carbon and catalysts.
F I G U R E 1 The reactions catalyzed by the ethylene-forming enzyme (EFE) from Pseudomonas syringae. (A) Representation of the two parallel EFE-catalyzed reactions (red arrows) showing the substrates, cofactors and products. (B) A simplified nonstoichiometric representation of the heterologous EFE reactions (red arrows) integrated as part of the metabolic network of the cyanobacterium Synechocystis sp PCC 6803 engineered to produce ethylene from 2-oxoglutarate. The metabolic intermediates discussed in the text are shown in black font, and the corresponding enzymes in blue font. Enzyme cofactors and side reactions have been omitted for clarity. Arg, arginine; ADP, adenosine diphosphate; CBB, Calvin-Benson-Bassham cycle; CoA, coenzyme A; P5C, 1-pyrroline-5-carboxylic acid; PEP, phosphoenolpyruvate; TCA, tricarboxylic acid 2 | BIOTECHNOLOGICAL SYSTEMS USING PHOTOAUTOTROPHIC CYANOBACTERIA AS ETHYLENE-PRODUCING HOSTS There are at least two biological enzyme-catalyzed processes that release ethylene, which offer alternatives for the development of novel biotechnological production systems. Plants and some algae produce ethylene as a hormone involved in the control of developmental processes and stress responses (Binder 2020;Ju et al. 2015).
This pathway proceeds in two consecutive steps from S-adenosyl-Lmethionine (SAM) via 1-aminocyclopropane-1-carboxylic acid (ACC) to ethylene, catalyzed by the enzymes ACC synthase (ACS) and ACC oxidase (ACO), respectively (Bleecker & Kende 2000;Houben & van de Poel 2019). Although the system has been linked with possible constraints including limited abundance of SAM or the release of cyanide as a side product, it has been shown to be applicable for heterologous ethylene production (Jindou et al. 2014) and offers a target for further engineering. The current review is focused on another strategy based on microbial ethylene forming enzyme (EFE). This enzyme is found in several plant pathogens including the Gram-negative bacterium Pseudomonas syringae, which in nature produce ethylene to interfere with the infected plant's normal ethylene signaling. This is expected to render the plant more susceptible to damage, which is advantageous for the invading pathogen (Weingart et al. 2001). The ethylene-forming reaction catalyzed by EFE uses the tricarboxylic acid (TCA) cycle intermediate 2-oxoglutarate (2-OG) as the primary substrate and is therefore tightly coupled to the cellular carbon/nitrogen metabolism, as discussed in detail in the following sections.
Heterologous expression of EFE has been applied to produce ethylene in several heterotrophic microorganisms including Escherichia coli (Fukuda et al. 1992a(Fukuda et al. , 1992bLynch et al. 2016) and Saccharomyces cerevisiae (Pirkov et al. 2008;Johansson et al. 2014;Johansson et al. 2017; see review Eckert et al. 2014). These systems are fundamentally restricted by the dependence on carbohydrate starting materials, which typically derive from relatively limited sources of biomass, thereby setting a limit to the upscale potential. An alternative strategy is to express EFE in photosynthetic cyanobacteria to achieve direct solar-driven production of ethylene from CO 2 (Eckert et al. 2014; Figure 1), which as a process is energetically much more favorable in comparison to any system that uses plant biomass as starting material.
The concept is founded on long-term biochemical research on cyanobacterial metabolism and molecular engineering, which has branched off towards the development of novel biotechnological applications (Angermayr et al. 2015;Hagemann & Hess 2018;Luan et al. 2020;Luan & Lu 2018). Because ethylene is a volatile gas, it requires a culture system with a closed gas circuit but enables endproduct harvest without the need for separation of biomass or extraction from the growth medium. In addition, since ethylene diffuses spontaneously out from the cell and into the culture headspace, potential problems with product toxicity and end-product inhibition are avoided. Although the overall approach is very attractive, there are still many biological and technological hurdles that need to be overcome in order to allow the transition towards efficient large scale production systems (Markham et al. 2016). For this to be possible, it is necessary to understand the function of the cyanobacterial host at molecular level, and to be able to extrapolate the current knowledge towards new metabolic engineering and cultivation strategies. This review focuses on the development of cyanobacterial ethylene production systems and provides a perspective to different aspects of integrating the heterologous EFE pathway as part of the cyanobacterial metabolic network.
The research on cyanobacterial ethylene production dates back to the end of the 1980s and is primarily based on the expression of heterologous EFE (Figure 1)  that shares sequence homology and overall fold with dioxygenases in the Fe(II)/2-OG oxygenase superfamily. Extensive molecular studies including biochemical functional characterization, 3D-crystallography, structure-based mutagenesis and kinetic analysis have provided detailed information on the complex EFE-catalyzed reaction, the catalytic mechanism and structure-function interactions (Martinez et al. 2017;Martinez & Hausinger 2016;Zhang et al. 2017). EFE is a unique enzyme both in the 3D fold and function, as it shares structural characteristics with both 2-OG oxygenase subgroups I and II, with a distinct active site topology, and essentially, the unusual ability to produce ethylene (Martinez et al. 2017).
EFE catalyzes a bifurcated reaction with two distinct conversion reactions that essentially compete with one another (Fukuda et al. 1992a(Fukuda et al. , 1992bMartinez & Hausinger 2016; Figure 1A). In the ethylene-forming reaction [EC 1.13.12.19], which is unique and of specific interest in this review, the enzyme catalyzes a four-electron oxidation of 2-OG into ethylene and three equivalents CO 2 . The reaction requires molecular oxygen as co-substrate, and is dependent on Fe(II) and L-Arginine (L-Arg) as cofactors. Notably, L-Arg is not consumed in this reaction (Martinez & Hausinger 2016), in contrast to T A B L E 1 List of engineered cyanobacterial ethylene production systems described in literature  (Durall et al. 2020) what has been anticipated earlier, which may have important metabolic implications for future engineering. In the alternative second reaction (EC 1.14.11.34) which is typical to 2-OG oxygenases, EFE catalyzes the C5-hydroxylation of L-Arg into 5-hydroxy-L-Arg, with concomitant oxidative decarboxylation of 2-OG into succinate. Subsequently, the resulting 5-hydroxy-L-Arg decomposes nonenzymatically into the two final products guanidine ) and L-Δ-1-pyrroline-5-carboxylic acid (P5C). The ethylene-forming reaction and the hydroxylation of L-Arg are thus separate events, and represent two alternative scenarios (Martinez & Hausinger 2016). The ratio between ethylene and succinate synthesis has been reported to be 2:1 (Fukuda et al. 1992a(Fukuda et al. , 1992bMartinez & Hausinger 2016), but this is likely not fixed, and depends on the exact reaction conditions. Combined with molecular structural information, this could provide an opportunity for enhancing ethylene production by attenuating the unwanted L-Arg hydroxylation by structure-based enzyme engineering, or by optimizing the process conditions. EFE is composed of 10 α-helices and two β-sheets that form a distorted double-stranded β-helix (DSBH) core, with an unusually hydrophobic active site pocket at the interface of the secondary structure elements. The enzyme undergoes extensive structural rearrangement upon L-Arg binding by induced fit mechanism (Zhang et al. 2017), which is essential for the catalysis, and involved in determining the reaction specificity between ethylene and succinate formation. In the proposed complex substrate recognition pattern, the 2-OG substrate is first coordinated to Fe(II) in the active site, followed by L-Arg binding, which results in conformational changes in the active site topology, and the alignment of the different reacting species (Martinez et al. 2017). The atypical ability to produce ethylene appears to be linked with changes in metal coordination, dioxygen binding site, and the hydrogen bonding pattern around the substrates.
Ultimately, the L-Arg in the active site is aligned in a configuration that does not favor C5-hydroxylation (off-line configuration) and enables the ethylene-forming reaction to take place. Consistent with alternative binding modes between the substrates, the ethyleneforming reaction shows substrate inhibition in regards to both L-Arg and 2-OG (Zhang et al. 2017), emphasizing the intramolecular factors dictating the outcome, and the complexity of the reaction when it comes to optimizing ethylene production. While this obviously reflects on the interpretation of the kinetic parameters, the apparent K M values for both 2-OG and L-Arg are in the micromolar range, whereas the apparent K cat for ethylene formation has been reported to be around 2 s À1 (Martinez & Hausinger 2016). Notably, the catalytic efficiency K cat /K M is significantly higher for ethylene formation in comparison to L-Arg hydroxylation (Martinez & Hausinger 2016), which further supports the view of the ethylene reaction being the predominant pathway in EFE.
The ethylene-forming reaction consumes equimolar amounts of 2-OG and is thus metabolically connected to a number of endogenous pathways branching off from the TCA cycle, including the biosynthesis of derived amino acids (glutamate, glutamine, proline in addition to L-Arg) and purine metabolism (guanidine). Importantly, as 2-OG provides the carbon skeleton for nitrogen assimilation and serves as an indicator for the intracellular carbon/nitrogen balance in cyanobacteria (Fokina et al. 2010), the use of this intermediate pool for ethylene biosynthesis may affect central metabolic flux rates and intracellular equilibria such as ATP/NADPH ratiowith potential downstream effects at the cellular level. Some of the associated key considerations are discussed in more detail in the following sections from the viewpoint of ethylene production and biotechnological limitations.
3.1 | Ethylene and EFE as the limiting factors in heterologous cyanobacterial production systems Ethylene in itself appears not to be toxic to cyanobacterial cells.
Adverse physiological effects that would restrict the development of This is reflected on the native carbon distribution in the cell, which may severely restrict the availability of 2-OG for ethylene production.
However, the cyanobacterial metabolic network appears to be relatively flexible and adjustable to changing environmental and genetic conditions , as seen, for example, in strains that are able to efficiently shift between different trophic modes depending on the availability of light, CO 2 and carbohydrate substrates. This plasticity is also reflected on our ability to tailor the strains ), yet while being beneficial, it also allows the host cell to In context with the high substrate demand, it should be noted that the EFE-catalyzed reaction releases three equivalents of CO 2 in each reaction cycle from five-carbon 2-OG into ethylene ( Figure 1A), which makes the overall pathway very carbon-inefficient (Eckert et al. 2014). Although this is mechanistically a wasteful reaction, the released CO 2 is destined to be recycled back to be fixed by RuBisCO in the CBB cycle, and is thus not lost from the system, unlike in heterotrophic host organisms that are unable to fix CO 2 . In this respect, the EFE pathway via 2-OG appears well compatible with the photoautotrophic machinery, although the true net impact of the CO 2 release cannot be quantitatively assessed before the dynamic interactions with the surrounding metabolism have been mapped out.

| Ethylene production via 2-OG in context with nitrogen metabolism in cyanobacteria
Besides the carbon flux, photoautotrophic ethylene production from 2-OG is intimately linked with cellular nitrogen metabolism and various regulators in the reaction network. At the precursor level, 2-OG is the substrate for glutamate, and therefore serves as the carbon skeleton for the glutamine synthetase-glutamate synthase (GS-GOGAT) cycle, which is the main ammonium assimilation pathway in cyanobacteria (see review Zhang et al. 2018). Restricted availability of 2-OG would therefore limit the process, with potential impact on ethylene production as well as the maintenance of native carbon/nitrogen homeostasis, that may cause many indirect secondary effects.
At regulatory level, 2-OG plays a key role as a signaling molecule indicating nitrogen starvation in the cell (Zhang et al. 2018). In this cascade, 2-OG is bound by NtcA, the global nitrogen regulator in cyanobacteria, which triggers various downstream cellular responses when nitrogen is scarce (Zhao et al. 2010). At this common interface between ethylene biosynthesis and regulation of nitrogen metabolism, the expression levels of NtcA have been shown to inversely correlate with ethylene production in Synechocystis strains expressing EFE (Mo et al. 2017). This is expected to be related at least in part to the increased flux towards glutamine biosynthesis via GS, as promoted by NtcA, which drains the flux away from 2-OG and EFE. In parallel, ethylene biosynthesis is likely to be affected by the global signal transduction protein P II that also regulates nitrogen metabolism in cyanobacteria (Scholl et al. 2020). P II is involved in sensing the overall energy and carbon/nitrogen status of the cell ( The introduced ethylene pathway is interconnected with the cellular carbon/nitrogen metabolism also via L-Arg, which is regulated in response to nitrogen availability. The biosynthesis of L-Arg is dependent on N-acetyl-L-glutamate kinase (NAGK), an enzyme catalyzing the committed step in cyclic arginine synthesis, that is regulated by the same P II control circuit that modulates the activity of NtcA and PEPC. When nitrogen is abundant and 2-OG levels are low, the system upregulates the production of L-Arg, and further, enhances the flux towards the cyanobacterial nitrogen reserve cyanophycin (multi-L-arginyl-poly-L-aspartic acid; Maheswaran et al. 2006).
Although L-Arg is required as a cofactor for ethylene synthesis and consumed in the competing branch of the EFE reaction, it is likely to limit ethylene production less than the availability of 2-OG. This is consistent with the observations that the supplementation of L-Arg in the culture medium has not had any clear positive effect on ethylene production in Synechococcus (Carbonell et al. 2019) or Synechocystis (Durall et al. 2020), while in heterotrophic Saccharomyces cerevisiae L-Arg addition actually reduced the yield (Johansson et al. 2014).
The intracellular signals associated with nitrogen limitation, whether caused by conditional factors or genetic modifications that increase the 2-OG levels, also link to the central carbon metabolism, and trigger the accumulation of glycogen reserves (Aikawa et al. 2012;Yoo et al. 2007). In this regulatory circuit, the buildup of 2-OG alters the crosstalk of P II with another two interactors, PirC (PIIinteracting regulator of carbon metabolism) and PGAM (2,3-phosphoglycerate-independent phosphoglycerate mutase), consequently redirecting the cellular carbon flux towards glycogen storage (Orthwein et al. 2021). In context with ethylene production, the The ethylene pathway also generates nitrogen-rich guanidine (CH 5 N 3 ) as a side-product in the competing EFE-catalyzed reaction via 5-hydroxy-L-Arg (Figure 1). As a toxic compound, guanidine must be removed either by degradative pathways that recycle nitrogen back into use, or by transport out from the cell (Nelson et al. 2017).

| Cultivation strategies for ethylene production in cyanobacteria
Besides the genetic background, the culture conditions have a significant effect on the production of ethylene by engineered cyanobacteria.
As ethylene biosynthesis via the EFE pathway is dependent on the metabolic flux down to the TCA cycle and 2-OG, conditional factors that alter the metabolic state of the host cell directly influence the amount of the available resources funneled to the target pathway. As photosynthetic organisms are not primarily dependent on the TCA cycle as a catabolic pathway for generating biological energy equivalents (ATP and NADPH), and mainly use it to obtain amino acid precursors for protein synthesis, the overall flux towards 2-OG under different trophic conditions (photomixotrophy and photoautotrophy) can be severely limited (Wan et al. 2017). The flux ratios together with the total efficiency of the light conversion and CO 2 fixation, which directly influence ethylene production, are therefore determined by external factors such as light, and the availability of carbon and nitrogen. These factors, as experienced by individual cells in the system, are dependent on the specific cultivation setup and equipment used for ethylene production and differ from one published study to another (Table 1).
In most published cases, ethylene production by cyanobacterial hosts has been studied in lab-scale liquid batch cultures in Erlenmeyer flasks (Table 1). In these systems, the aeration, and hence the gas transfer between the headspace and the medium, is facilitated simply by mixing. This has a direct impact on the distribution of CO 2 to the cells, which typically limits optimal photoautotrophic production. In more advanced photobioreactor systems, gas may be actively pumped into the medium, which can improve the availability of carbon in the culture. The carbon dioxide concentrations typically vary between ambient and 5% CO 2 in the gas phase, while bicarbonate is in many cases supplemented as an additional source of inorganic carbon. In this context, also the temperature (20-30 C), the medium pH (7.5-8) and buffer capacity affect the solubility and hence the carbon uptake efficiency. While higher CO 2 concentrations boost the cellular carbon uptake in general, the impact on ethylene production is not necessarily linear, and dependent on combined conditional and genetic effects.
The carbon partitioning is affected by the overall energy status of the cell, and increased CO 2 availability may promote biomass accumulation, the flux towards storage compounds, or the excretion of organic acids through overflow metabolism at the expense of ethylene, if the system is not in balance (Cano et al. 2018;Gründel et al. 2012). It is important also to note that the availability of oxygen, which is essential for the EFE-catalyzed reaction as well as for the oxidative TCA cycle, is likely to be less restricted in oxygen-producing photoautotrophic cyanobacteria as compared to heterotrophic production systems, in which ethylene production is clearly limited by the O 2 solubility and transfer through the medium (Johansson et al. 2013).
The performance of photoautotrophic production systems is also critically affected by the light conditions. Apart from the incident light source that determines the quality and quantity of incoming radiation, the distribution of light to individual cells is the sum of many parameters. These include cultivation vial dimensions, the length of the light path, culture optical density and mixing, that consequently influence the level of absorption, self-shading and light fluctuation. As in most cases in cyanobacterial research in general, ethylene production has been studied mainly under constant white light. The use of moderate growth light appears to promote ethylene production over higher light intensities (Ungerer et al. 2012). This may reflect the level of direct photodamage and ROS-induced stress, and the allocation of resources for maintenance functions that restricts the optimal use of cellular resources for ethylene biosynthesis under increased illumination, but also possible direct effects on EFE expression. In comparison to suspension cultures, the light conditions can be more effectively controlled in respect to temporal fluctuations and homogeneity in solid-state culture systems where the cells are static and entrapped in thin layer polymer matrix (Vajravel et al. 2020).
In addition to CO 2 and nitrogen, the chemical composition of the rest of the culture medium (Table 1) affects cell growth and the overall productivity. While most cyanobacterial ethylene production systems are based on BG11 medium, the use of concentrated medium has been shown to increase productivity (Ungerer et al. 2012;Zhu et al. 2015). As no single component could be attributed to the effect, the limitation appears to be a sum of multiple nutrient-associated factors (Zhu et al. 2015), which are not directly related to carbon or nitrogen. Although most production strategies rely on culture conditions under which the cells operate in fully photoautotrophic mode, the effect of supplemented carbohydrate substrates that promote mixotrophic growth has also been studied. Generally, it appears that the availability of sugars in the cyanobacterial cultures enhances ethylene production. However, the effect is likely associated with faster growth and resulting increase in the total biomass in the culture, rather than improvement at the level of single cells resulting from increased respiratory flux towards the TCA cycle (Lee et al. 2015 (Vajravel et al. 2020). As growth in this setup is suppressed, possible constraints associated with nitrogen availability and regulation via 2-OG are likely less critical in comparison to actively dividing cells.

| Quantitative monitoring of ethylene production
Ethylene is a highly volatile compound that spontaneously diffuses out from the producer cell, and eventually into the culture headspace.
This alleviates the need for chemical extraction from the culture medium, which is beneficial for the development of a continuous bioproduction process but requires a closed system for product collection and isolation. The most conventional strategy used for quantitative ethylene analysis is to transfer an aliquot of the producer culture into a sealed small-scale vial, followed by incubation and the analysis of the gas-phase, typically using a gas chromatograph fitted with a flame ionization detector (GC-FID). It is noteworthy that this type of a two-phase procedure does not necessarily give precise information on the actual main culture, and that the exact analytical setup significantly affects the absolute calculated productivities, which together complicate meaningful quantitative comparisons between published studies. A more advanced system combines cultivation in a photobioreactor with a membrane-inlet mass-spectrometer (MIMS), which enables high-resolution analysis of ethylene (Zavřel et al. 2016) directly from the headspace. This type of a strategy may aid process parameter optimization during system upscale, as it allows real-time monitoring of long-term continuous production, and would be convenient for large scale industrial processes. The progress towards biotechnological photoautotrophic platforms to convert CO 2 directly into ethylene requires the development of continuous systems for step-wise upscale and process optimization.
In this transition from research laboratories to industry, the prospects of utilizing both suspension cultures as well as solid-state cultures need to be further systematically explored. Besides allowing more accurate evaluation of the system performance when the conditions are not in constant change, continuous production strategies also enable the fine-tuning of individual culture parameters for different strains. In addition, the use of direct analytical methods, such as MIMS or commercial probes for quantitating volatile organic compounds, would give access to more uniform data, more reliable comparative analysis, as well as system automatization, when compared to strategies that rely on separate analytical cultivation steps.
As for the biology, the current molecular-level knowledge on cyanobacterial ethylene production, combined with access to advanced synthetic biology tools, provides means for developing next-generation EFE expression systems with optimized gene dosage, transcription and translation. Although it is difficult to estimate how much the EFE expression levels ultimately need to be pushed up, as more effective strain-specific engineering strategies emerge to steer the cellular carbon flux towards 2-OG, the improvement of EFE catalytic performance and the substrate availability are expected to proceed hand in hand. Due to the intimate role of 2-OG and L-Arg in cellular carbon/nitrogen metabolism, this will need to take into account the host organism as a whole, including the complex regulatory networks that govern the biosynthetic and catabolic equilibria. In order to achieve sufficient photon conversion efficiencies, we must fuse the ethylene-specific engineering steps with different strategies to increase the overall photosynthetic carbon fixation with flux analysis and metabolomics to pinpoint the next metabolic bottlenecks. In the light of the current knowledge, the potential needs to be evaluated together with EFE side products such as guanidine, which need to be effectively recycled back to use, unless extracted in parallel for added commercial value. For all this, profound exploitation of computer-based modeling and bioinformatics will be necessary to interpret the complex data, and to design and evaluate further metabolic modifications.
There is room to critically improve the photoautotrophic production of carbon-based target chemicals such as ethylene in engineered cyanobacterial hosts. Reaching these novel industrial solutions, however, will require resources and persistent long-term multidisciplinary collaboration across different fields of fundamental biological research, applied biosciences, and industrial technology development.

ACKNOWLEDGMENTS
This review article was prepared as part of the collaboration in the Nordic Centre of Excellence (NCoE) NordAqua funded by NordForsk through the Nordic Bioeconomy Programme (project #82845). The authors have no conflict of interest to declare.

AUTHOR CONTRIBUTIONS
All authors contributed to the planning, writing and revision of the manuscript.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.