An adjustable algal chloroplast plug-and-play model for genome-scale metabolic models.

The chloroplast is a central part of plant cells, as this is the organelle where the photosynthesis, fixation of inorganic carbon, and other key functions related to fatty acid synthesis and amino acid synthesis occur. Since this organelle should be an integral part of any genome-scale metabolic model for a microalgae or a higher plant, it is of great interest to generate a detailed and standardized chloroplast model. Additionally, we see the need for a novel type of sub-model template, or organelle model, which could be incorporated into a larger, less specific genome-scale metabolic model, while allowing for minor differences between chloroplast-containing organisms. The result of this work is the very first standardized chloroplast model, iGR774, consisting of 788 reactions, 764 metabolites, and 774 genes. The model is currently able to run in three different modes, mimicking the chloroplast metabolism of three photosynthetic microalgae-Nannochloropsis gaditana, Chlamydomonas reinhardtii and Phaeodactylum tricornutum. In addition to developing the chloroplast metabolic network reconstruction, we have developed multiple software tools for working with this novel type of sub-model in the COBRA Toolbox for MATLAB, including tools for connecting the chloroplast model to a genome-scale metabolic reconstruction in need of a chloroplast, for switching the model between running in different organism modes, and for expanding it by introducing more reactions either related to one of the current organisms included in the model, or to a new organism.

Introduction

The chloroplast is a vital part of any plant cell. This is the organelle where the light reactions of photosynthesis occur [1-3], and also the organelle that hosts the Calvin-Benson cycle [2,4,5]. Since the Calvin-Benson cycle is the single most important pathway for fixation of inorganic carbon, the chloroplast is responsible for nearly 100% of the primary production carried out by plants, algae and cyanobacteria. This particular organelle also contains important pathways related to fatty acid biosynthesis [6,7], pigment synthesis [8,9] and de novo synthesis of several amino acids [10,11] which are of great importance to mammals, as we are not able to produce most of these ourselves [12,13].

As the chloroplast is an essential part of a plant cell, having a good mathematical representation of the chloroplast is also essential to any model describing the metabolism of a plant or algal cell.

Since DNA sequencing technology has seen great improvements in the later years, and new genomes are being sequenced at an increasing pace, the need to ease construction of metabolic models have also emerged. As a result of this, automated tools for genome annotation and model construction have been developed [14,15]. This has improved the time aspect of creating metabolic models for newly sequenced organisms, and also contributed to a standardization of the construction of metabolic models. However, some of the automatic model construction approaches that have been developed, do not provide the user with detailed insight into how the model is being constructed or how gaps in a draft metabolic network are filled. Since these tools are automated and purely based on the information available in databases, pathways and reactions that are not actually present in the organism might be introduced to the model, while other reactions might be missing if the enzymes responsible for catalysing them are organism-specific and not yet present in databases. As of yet, metabolic models assembled by an automated approach also give the user few opportunities to alter the resulting model.

As an alternative approach to ease model construction of chloroplast-containing cells, we have developed a standardized chloroplast model as a plug-and-play module that can be run in different organism modes. The chloroplast model is built on chloroplast-specific pathways found in other models of chloroplast metabolism, but the photosynthetic processes, which are often incorrectly represented in published metabolic models for photosynthetic organisms, have been carefully curated in this chloroplast model. This novel approach of developing an organelle-focused plug-and-play module that standardizes the relatively conserved chloroplast metabolism, also allows for the fine-tuning of minor metabolic differences between different organisms containing a chloroplast. Thus, we propose the approach of manually generating high-quality organelle template genome-scale metabolic modules as a new paradigm for improved automated reconstruction of eukaryotic genome-scale metabolic models. Additionally, we have developed several computational tools for connecting the chloroplast module to a pre-existing genome-scale metabolic model lacking a chloroplast, for changing organism-mode of the chloroplast, and for making it easy to introduce additional reactions to the model in general or to the organism-specific parts of it. The software allows new organism modes to be introduced.

Materials and methods

Reconstruction of the draft metabolic model

Previously published reconstructions of Nannochloropsis gaditana [16], Chlamydomonas reinhardtii [17] and Phaeodactylum tricornutum [18] were used as basis for our draft chloroplast model iGR774. The N. gaditana model was used as a base for most of the common chloroplast metabolic processes shared between the three algae, while the C. reinhardtii model and the P. tricornutum models were used for introducing organism-specific parts as a proof-of-principle.

To be able to distinguish between reactions common to all three organisms and strictly organism-specific reactions, the organism-specific reactions were given a tag indicating which organism the reaction in question is specific to. For Nannochloropsis, the organism specific tag is ‘@Nan’, which was used as a prefix to the ID of the reaction in question (example: '@Nan_R05345_h'). For Chlamydomonas and Phaeodactylum specific reactions, ‘@Chl’ and ‘@Pha’ were used as organism specific reaction tags. These tags are used by the script developed to switch the model from one organism-mode to another in order to recognize organism-specific reactions, and determining the flux limits for these in the current organism-mode.

Curation of the draft metabolic model

Since the aim of this work was to create a standardized model of chloroplast metabolism, the draft metabolic model was curated by standardizing pathways known from literature to be present in the chloroplasts of Nannochloropsis, Chlamydomonas and Phaeodactylum. Special emphasis was put on pathways involved in lipid synthesis, pigment synthesis and synthesis of animo acids, as these are groups of molecules that are currently attracting much interest from the academic community. Other pathways, for example the ones involved in synthesis of RNA and DNA, were omnitted, as these are not considered essential in the function of a metabolic chloroplast module. Certain reactions also needed to be introduced during reconstruction, and for this purpose, tools were developed to ease the import of new reactions to the model from KEGG (Kyoto Encyclopedia of Genes and Genomes) (see section on method development).

Since photosynthesis is an essential part of the chloroplast and its normal function, the photosynthetic electron transfers were also modelled in detail as part of the curation process. When modelling photosynthesis as part of a metabolic model of a photosynthetic organism, it has been common to describe each photosynthetic complex as one distinct reaction. In some reaction system reconstructions, additional processes related to photosynthetic activity, such as cyclic electron transport [19-20], the Mehler reaction [21] and plastoquinone re-oxidation by the enzyme Plastid Terminal Oxidase (PTOX) [22] has also been modelled as part of the photosynthetic activity.

To give a more detailed overview of the processes involved in photosynthesis, and to be able to assess electron fluxes, we modelled every electron transfer process occurring in the photosynthetic electron transport chain as one distinct reaction, yielding 33 photosynthetic reactions, instead of the usual 5 to 8.

Genetic information was also added to the chloroplast model (see the genes section for more information). Finally, as part of the curation process, the maximum and minimum fluxes of certain reactions were restricted (for a complete list of restricted reactions, see S1 Table). This was necessary to make the model produce a realistic set of metabolic fluxes, as the chloroplast model is a module meant to be fused with an exo-model. Since the chloroplast module is developed to be a part of another model in need of a chloroplast, the chloroplast module includes a set of cytosolic metabolites, in addition to several exchange reactions responsible for transporting these metabolites between the chloroplast and the cytoplasm. When the chloroplast module is coupled to an exo-model, the script being responsible for the fusion of the models will localize the exo-model version of the exchange metabolites of the chloroplast, and the names of the chloroplast exchange metabolites will be changed into the namespace of the exo-model, to ensure a seamless fusion of the two models. If one or more of the chloroplast exchange metabolites are not present in the exo-model, the user will be notified to aid in the debugging in case the combined model does not run. When the chloroplast model operates on its own, no cytosolic reactions are present to generate or consume the exchanged metabolites. When run as a standalone model, the cytosolic metabolites are therefore allowed to be imported or exported freely from the model’s external environment. These reactions (tagged ‘B_’ for ‘border reaction’ in the chloroplast model) are automatically deleted when the chloroplast model us used as a module that is fused with an exo-model.

Many exchange reactions are involved in exchange of carbon compounds. If these exchange reactions were allowed to be run in both directions without constraints, the chloroplast model, when tested on its own, would be perfectly able to produce / import all the compounds it is supposed to produce, without running photosynthesis and photosynthetic carbon fixation.

Genes

The chloroplast model was supplied with genetic information, mainly from KEGG and NCBI (National Center for Biotechnology Information). KEGG reactions are sometimes accompanied by information about the protein catalysing them in different organisms. When a KEGG reaction was imported to one of the organism-specific parts of the model by the script pickKEGGrx (described in S1 Text), the KEGG API (Application Programming Interface) and MATLAB (Matrix Laboratory, MathWorks, USA) was used to search for the EC (Enzyme Commission) number of the enzyme(s) catalysing the KEGG reaction in question, and for each KEGG number found, KEGG API and MATLAB was again used to search for Nannochloropsis, Chlamydomonas or Phaeodactylum specific genes catalysing the enzyme in question. If genes were found, they were added to the chloroplast model for the reaction in question.

For adding genetic information to the photosynthetic electron transfers carried out by one of the three multisubunit protein complexes of the photosynthetic electron transport chain; Photosystem II (PSII), the cytochrome b6f complex (b6f) or Photosystem I (PSI), or the reaction carried out by the proton-driven complex ATP synthase, the NCBI Protein database was used. A list of all gene–protein associations in Nannochloropsis was downloaded, and the list was scanned for the protein subunits of PSII, the cytochrome bf complex, PSI or ATP synthase. Genes for several subunits of all four protein complexes were identified. These were added to the model with an ‘AND’ relationship for all electron transfers occurring in the complex in question.

Chloroplast biomass reaction

When the chloroplast model is merged with an exo-model, the exo-model’s biomass reaction will be the prevailing biomass reaction, and might be modified manually by the user to account for the addition of a chloroplast.

However, to be able to test the chloroplast model’s ability to produce the many important metabolites it should be able to produce, a separate biomass reaction had to be added to the chloroplast model. This ‘chloroplast biomass’ reaction is deleted by the script plugging the chloroplast model into an exo-model, but when the chloroplast model is tested on its own, this is the reaction that will be optimized.

The most important metabolite classes produced by the chloroplast are the membrane glycolipids monogalactosyldiacylglycerol (MGDG), digalactosyldiacylglycerol (DGDG) and sulfoquinovosyldiacylglycerol (SQDG) [11], the amino acids lysine, methionine, threonine, phenylalanine, tyrosine, tryptophan, leucine, isoleucine and valine [11,13], an organism-specific pool of pigments [8,23] in addition to ATP and NADPH, which are the end-products of the photosynthetic electron transport chain.

The chloroplast specific biomass reaction is therefore composed of the chemical equations ATP + H2O → ADP + Pi (Pi denoting inorganic phosphate), NADPH → 2 e−+ NADP+ + H+, and O2 + 4 H+ + 4 e−→ 2 H2O, the latter equation being present in the biomass solution in order to balance out the electrons between the left and the right side of the equation. The organism-specific chloroplast-produced fractions of membrane lipids, proteins and pigments are also consumed in the chloroplast biomass reaction. Keeping electron balance in mind, and accounting for the fact that the chloroplast usually produces more ATP than NADPH, the following reaction was used as chloroplast biomass:

In the chloroplast biomass equation above, CPr denotes chloroplast proteins, CML denotes chloroplast membrane lipids and CPi denotes chloroplast pigments.

The composition of the chloroplast protein fraction, the membrane lipid fraction and the pigment fraction consumed in the chloroplast biomass reaction are organism-specific, and in practice, one organism-specific biomass reaction is present for every organism-mode included in the chloroplast model. Currently, one biomass reaction is present for each of the three organisms Nannochloropsis, Chlamydomonas and Phaeodactylum, but the chloroplast model can be expanded to be able to describe the metabolism of more organisms, and for every new organism mode that is added, one biomass reaction must also be added to describe the chloroplast production of the organism in question. When switching between organism modes, the biomass reaction is also changed accordingly, unless the chloroplast model is coupled to an exo-model. Then the biomass reaction of the exo-model will be the reaction to be optimized.

The organism-specific biomass components accounting for membrane lipids, protein and pigment production were modelled according to stoichiometric values found in literature. For Nannochloropsis, the ratios between the different fatty acids in different membrane lipids were taken from Vieler et al. 2012 [11], while the ratio between MGDG, DGDG and SQDG in the cell membrane was taken from Li et al. 2014 [24]. The ratios between different fatty acids in the different classes of membrane lipids in Chlamydomonas were taken from Suh et al. 2015 [25], while the ratio between the different lipid classes was taken from Boudière et al. 2014 [26].

In Phaeodactylum, the ratios between fatty acids in the different membrane lipids were taken from Tonon et al. 2002 [27], and the ratio between different membrane lipids was found in Arao et al. 1987 [28].

The mean composition of the different amino acids in an average protein was taken from Xiao et al. 2013 [29] for Nannochloropsis, from Boyle & Morgan, 2009 [30] for Chlamydomonas and from Brown, 1991 [31] for Phaeodactylum.

The pigment composition in Nannochloropsis was found in Lubián et al. 2000 [8], and in Eichenberger et al. 1986 [23] for Chlamydomonas. Phaeodactylum contains some pigments that are not found in the two other organisms, for example fucoxanthin, diatoxanthin, diadinoxanthin and chlorophyll c in addition to chlorophyll a [32]. The production of these require specific pathways that could be added to the model in the future. The detailed composition of the chloroplast biomass components ‘chloroplast proteins’, ‘chloroplast membrane lipids’ and ‘chloroplast pigments’ is described in S2 Text.

Constraint-based linear optimizations

Mathematically, the metabolic network of iGR774 is represented by an M x N sized stoichiometric matrix S, where M denotes the number of metabolites, and N denotes the number of reactions. A positive stoichiometric coefficient s, thus indicate that s molecules of the ith metabolite is produced in the jth reaction. A negative stoichiometric coefficient indicates consumption of the metabolite instead of production. Flux balance analysis (FBA) was used for optimization of the model under criteria of steady-state, represented by the linear problem where ν is the flux vector that best solves the optimization problem.

For a list of reactions that were restricted, and the reason for restricting the reaction in question, see S1 Table.

FBA optimizations were performed using MATLAB and COBRA Toolbox [33] in combination with the Gurobi optimizer (version 8.1.0, Gurobi Optimization Inc., Houston, Texas).

Method development

The development of a sub-model meant to be plugged into an exo-model, and its ability to run in different organism-modes, is to our knowledge a new concept. For this reason, the model reconstruction work was accompanied by the development of a new toolbox of scripts. The newly developed scripts were written in MATLAB code, and will ease the process of working with a compartment model with the ability to be connected to another model.

When constructing the model, certain pathways specific to one of the organisms included in the chloroplast model had to be transferred from other models. An example of such a pathway was the eyespot which was transferred from the Chlamydomonas model published by Imam et al. in 2015 [17]. A specific script was written for the purpose of transferring reactions. The script simultaneously translate the metabolites into the namespace of the exo-model, and also checks that the reaction names are not already present in the chloroplast model.

Certain reactions were introduced to the chloroplast model from KEGG. Another script was therefore written to enable the import of KEGG reactions to the model. This script also translates the metabolites of the reaction in question to the chloroplast model namespace.

The chloroplast model currently contains 788 reactions, but it could still be further expanded, both by adding reactions to the organism modes already included in the model and by including new organism modes. The scripts for adding reactions, both from other models, from KEGG and with inputs from the user could ease the future process of expansion of the chloroplast model.

In order to ease the process of making the model more detailed, scripts were also developed for adding genetic information to the model, both to existing reactions, and to new ones.

In order to identify gaps in pathways, or to verify that a certain pathway (or the entire model, for that matter) is connected, several tools were developed. One script allows the user to follow a path of metabolites from a certain metabolite, treating the metabolite in question as either a substrate or a product. In order to check if a pathway or an entire model is connected, scripts were also written to transfer a model or part of a model to a format that can be imported into cytoscape, enabling the user to inspect the model visually (metabolic models can also be visualized and analysed by the newly published tool ModelExplorer [34]).

Scripts were also written to plug the chloroplast model into an exo-model, and to change organism-mode of the chloroplast model.

When plugging the chloroplast model into an exo-model, the metabolite names and IDs of the chloroplast metabolites are first translated to the namespace used by the exo-model, and the basis for comparison between the metabolites of the two models are KEGG IDs. Several models do, however, not contain KEGG IDs for their metabolites, and this problem we solved by writing two scripts for supplying a model with metabolite KEGG IDs. Since a metabolic model can include several thousand metabolites, one of these scripts is running in a very automatic manner, requiring little input from the user, while the other requires more input from the user, and is meant to complement the first script.

Last but not least, several general scripts were written aiming at making it easier to work with a metabolic model in matlab, and also to display information about enzymes, metabolites and reactions from KEGG in MATLAB.

To access the newly developed tools, see S2 File, and for a detailed description of the tools, see S1 Text.

Results and discussion

Summary of reconstructed chloroplast model

The result of the reconstruction of chloroplast metabolism was the iGR744 model (S1 File) containing 788 reactions, 764 metabolites and 774 genes. The reactions of the model belong to one or more of 71 subsystems (the number of reactions affiliated with the most important overall subsystems groups is shown in Fig 1).

Fig 1

The number of chloroplast reactions affiliated with each major subsystem group.

There were 110 genes that were shown to be critical for biomass production. Knockouts of 107 of these genes completely restricted the chloroplast model from producing biomass, while knockouts of 3 genes decreased the rate of biomass production. A list of the critical genes can be found in S2 Table.

There were 702 of the reactions included in the model that proved to be insensitive to organism mode, while 86 reactions were shown to be organism-specific.

There are 16 Nannochloropsis specific reactions in the model. These are mainly connected to synthesis of pigments that are only present in Nannochloropsis, and not in the other two organisms contained in the model. There are 50 reactions that are Chlamydomonas specific. Most of these are reactions occurring in the Chlamydomonas eyespot [35], which is an organelle Nannochloropsis and Phaeodactylum do not possess. There are also 16 reactions that are Phaeodactylum specific, and four organism-specific reactions are shared between Phaeodactylum and Chlamydomonas, as both of these organisms have the possibility of producing the fatty acid docosahexaenoic acid (DHA), although in relatively small quantities [25]. Chlamydomonas in addition produces C18:3, while this lipid is not produced to a large extent in Nannochloropsis and Phaeodactylum [11, 25, 27]. Each specific organism mode also includes reactions for generating MGDG, DGDG, SQDG, proteins and pigments. In addition, MGDG, DGDG and SQDG are combined into an organism-specific chloroplast membrane lipid component, where the different membrane lipids are added according to the correct organism-specific ratio. The exact composition of the organism-specific biomass components is given in S2 Text.

The chloroplast model is able to produce biomass running in both Nannochloropsis mode, Chlamydomonas mode, and Phaeodactylum mode. The growth rates are shown in Table 1.

Table 1

Growth rates of the chloroplast model simulated in Nannochloropsis mode, Chlamydomonas mode and Phaeodactylum mode, respectively.

Organism mode	Growth rate [mmol gDW-1 h-1]
Nannochloropsis	0.0318
Chlamydomonas	0.0315
Phaeodactylum	0.0455

As Table 1 show, the chloroplast growth rates are not the same when running in different organism modes. This is due to the fact that all the three possible organism modes currently introduced to the model have their own biomass reaction, adjusted to the chloroplast metabolism of the organism in question. The growth rates are similar when running the chloroplast in Nannochloropsis mode and Chlamydomonas mode, while the growth rate is higher when the chloroplast is run in Phaeodactylum mode. The increased growth rate for the Phaeodactylum mode might be due to the fact that Phaeodactylum has a specific pool of pigments, while the pathways responsible for producing these pigments have not yet been added to the model. The pigment pool included in the Phaeodactylum biomass reaction thus only contains chlorophyll a, which puts less requirements on the production pathways of the chloroplast compared to when it is running in Nannochloropsis or Chlamydomonas mode.

Photosynthesis

In metabolic models describing photosynthetic organisms, photosynthesis itself is usually modelled using very few reactions. The photosynthetic machinery consists of the three electron transferring complexes photosystem II, the cytochrome b6f complex and photosystem I. These transfer electrons, which eventually are used to regenerate NADPH from NADP+ by the enzyme Ferredoxin NADP+ reductase (FNR) after increasing the energy of the electrons using sunlight. Photosynthetic electron transport does not only lead to the generation of NADPH. It also pumps protons across the thylakoid membrane of the chloroplast. When photosynthesis is running normally, the concentration of protons in the thylakoid lumen is higher than in the stroma, due to PSII and the cytochrome b6f complex. This imbalance in proton concentration on the two sides of the thylakoid membrane is the driving force for ATP synthesis by the fourth protein complex, ATP synthase.

When creating metabolic models describing photosynthetic algae, it has been common to describe each electron-transporting complex as one reaction. In addition, some models contain reactions describing processes involved in regulating photosynthesis, such as cyclic electron transport [36], plastoquinone oxidation by Plastid Terminal Oxidase (PTOX) [37] and the Mehler reaction [38].

Photosynthesis is usually described by at most 10 reactions (The Phaeodactylum tricornutum model by Levering et al. published in 2016 [18] is an example). In several models (for example the Chlamydomonas reinhardtii model by Imam et al. published in 2015 [17], the Nannochloropsis gaditana model by Shah et al. published in 2017 [16], and the Nannochloropsis salina model by Loira et al. published in 2017 [39]), the reactions describing the electron transferring complexes in photosynthesis are in addition not correct with regard to proton pumping across the thylakoid membrane, which will affect the ratio of ATP-production to NADPH production.

The one-reaction-per-complex way of modelling photosynthesis makes it a black box. We therefore modelled every electron transfer occurring in the three major photosynthetic protein complexes, PSII, PSI and the b6f complex, making the electron-transporting fluxes more transparent. 10 of our modelled electron transfer reactions occur within PSII, 7 occur within the cytochrome b6f complex, 8 within PSI, while 3 are related to re-generation of NADPH by FNR. One reaction describes the re-generation of ATP from ADP by ATP synthase, two reactions are related to plastoquinone transport, while one last photosynthetic reaction describes cyclic electron transport around PSI, which uncouples the ratio between ATP and NADPH. The reconstruction of photosynthesis is shown in Fig 2.

Fig 2

Representation of photosynthesis in the chloroplast model.

The areas marked PSII, b6f, PSI and ATP synthase shows where in the photosynthetic electron transport chain the different electron transfers and reactions are taking place. Light blue nodes represent metabolites, while purple nodes represent reactions.

The NADPH and ATP molecules generated as a result of the photosynthetic electron transport are largely used by the cell to fix carbon in the Calvin-Benson cycle [2,4,5]. An interesting property of a chloroplast model with a special focus on photosynthesis would therefore be how the relationship between photon usage by photosynthesis and carbon uptake rate in the Calvin-Benson cycle affects cell growth.

A phenotype phase plane showing this relationship is shown in Fig 3.

Fig 3

Phenotype phase plane showing how the fluxes of carbon fixation (NanoG0589) and photon usage by PSII (PSII_photon) affect the flux of the chloroplast biomass function.

When optimizing the production of chloroplast biomass, the optimal flux through the RuBisCO-driven reaction responsible for CO2 fixation, the general import reaction for photons used by photosynthesis, and for the reaction partitioning photons to PSII are 125 mmol gDW-1 h-1 (DW represents dry weight), 1000 mmol gDW-1 h-1 and 500 mmol gDW-1 h-1, respectively. When running the chloroplast model, all photosynthetic reactions runs at their maximum allowed flux in all organism modes, restricted by the upper limit of photon import, which is set to 1000 mmol gDW-1 h-1.

The RuBisCO-driven reaction responsible for fixing CO2 is part of the Calvin-Benson cycle. This particular pathway is costly for the cells to keep running energy-wise, which is why it is functionally coupled to the photosynthetic electron transport chain.

When the Calvin-Benson cycle is running normally, it needs 3 molecules of CO2, 9 molecules of ATP and 6 molecules of NADPH to produce one single molecule of glyceraldehyde-3-phosphate or dihydroxyacetonephosphate. This ratio of demand for ATP and NADPH is fixed in the Calvin-Benson cycle, but the chloroplast’s need for the two energy carriers in general is not, as the general metabolism can be influenced by factors beyond the cell. The chloroplast therefore needs to be able to tune the production ratio of ATP to NADPH. This is done by performing cyclic electron transport around PSI [36]. Cyclic electron transport, as the name implies, sends electrons on a cyclic journey around PSI. When an electron is ‘recycled’ in this manner, they are sent back to the plastoquinone pool, which is a pool of mobile electron carriers shuttling between PSII and the cytochrome bf complex. Both the plastoquinone transport and the mechanism of the b6f complex contribute to the pumping of protons from the stroma into the thylakoid lumen. Cyclic electron transport thus adds to the proton gradient over the thylakoid membrane, while temporarily keeping electrons from being used to re-generate NADPH. Cyclic electron transport is thus the photosynthetic electron transport chain’s way of tuning the ATP:NADPH production ratio.

In the biological world, rate of cyclic electron transport is mediated by certain protein complexes [40], which are regulated by transcription factors; and by the partitioning of light-energy between PSII and PSI [41]. In optimizations of metabolism by FBA, however, the only regulating force is mathematics, and finding the most beneficial set of fluxes solving the linear problem in question. A mathematical model of metabolism might therefore occasionally take shortcuts. This proved to be the case in photosynthesis. Mathematically speaking, it seems to be beneficial for the model to optimize the photosynthetic ATP production, while using non-photosynthetic reactions to create NADPH. When the upper flux limit for the reaction carrying out cyclic electron transport is allowed to run unlimited, the optimal flux is 500 mmol gDW-1 h-1, which corresponds to a recycling of all photosynthetically generated electrons.

The motivation for building metabolic models, is to describe the metabolism of an actual cell. Even though it seems to be beneficial mathematically speaking to run cyclic electron transport at full speed, it does not describe what goes on in an actual cell, and the rate of cyclic electron transport have therefore been restricted with regard to the ATP and NADPH demands of the Calvin-Benson cycle.

To find the upper bound for cyclic electron transport best matching the demands of the Calvin-Benson cycle, a mini-model was built, consisting only of the photosynthetic electron transport chain and the Calvin-Benson cycle. This mini-model was optimized for production of dihydroxyacetonephosphate, and since it did not contain transport reactions for ATP, ADP, inorganic phosphate, NADPH or NADP+, the photosynthetic electron transport chain was forced to produce ATP and NADPH in the exact ratio needed for fuelling the Calvin-Benson cycle. The set of fluxes resulting from this optimization can be found in S3 Table.

The mini-model showed that the photosynthetic electron transport chain produces the exact ATP:NADPH ratio needed by the Calvin-Benson cycle, without having to run cyclic electron transport. This process have therefore been completely restricted in the chloroplast model, but the upper bound of cyclic electron transport can of course be changed for the purpose of analysing the impact of cyclic electron transport on varous aspects of the chloroplast metabolism.

Lipid synthesis

Since especially Nannochloropsis is a potential target for industrial lipid production, the ability of the Nannochloropsis chloroplast to produce lipids was explored. Even though specific classes of lipids might be of interest when using a microalga as a commercial lipid-producer, the chloroplast model contains a combined pool of important chloroplast-produced lipids, and when exploring the chloroplast model’s ability to produce Nannochloropsis-specific lipids, the production of this combined lipid pool was optimized.

Metabolism is a continuous process, and all the reactions are running simultaneously. Even so, certain reactions depend on the presence of other reactions to run. This is the case with lipid synthesis, which depends on carbon fixation by the Calvin-Benson cycle, which again depends on photosynthesis to fill its requirement for ATP and NADPH. The ability to produce lipids was therefore explored, using carbon fixation by RuBisCO and photon usage by PSII as control reactions. The result can be seen in Fig 4.

Fig 4

Phenotype phaseplane showing the dependency of lipid synthesis on photosynthesis (photon usage by PSII used as control reaction) and the Calvin-Benson cycle (CO2 fixation by RuBisCO used as control reaction).

Lipids could be produced as a maximum growth rate of 0.0819 mmol gDW-1 h-1. The corresponding fluxes for photon usage by PSII and CO2 fixation by RuBisCO were found to be 500 mmol gDW-1 h-1 and 150.038 mmol gDW-1 h-1, respectively. Furthermore, the phenotype phaseplane presented in Fig 4 nicely shows the lipid production’s dependency on both photosynthesis and the Calvin-Benson cycle, and the cycle’s dependency on photosynthesis, while photosynthesis is not dependent on either the Calvin-Benson cycle nor lipid production.

Conclusions

We have constructed the chloroplast metabolic model iGR774, containing 788 reactions, 764 metabolites, and 774 genes. This is the first metabolic model developed as a plug-and-play module to represent a high-quality reconstruction of the metabolism of a single organelle, and the module may be run in different organism modes. The iGR774 model is thus intended to be incorporated into another genome-scale metabolic model in need of a chloroplast organelle. We suggest the plug-and-play concept as a new paradigm for generating high-quality genome-scale metabolic models for eukaryotic organisms. Here, a general challenge in generating automated model reconstructions is associated with reaction localization to the different subcellular compartments. By instead developing a set of high-quality organelle modules, we propose that automated reconstruction frameworks may significantly improve their fidelity by using the modules as template reaction sets. A conceptually related approach to the use of template reaction sets is seen in the differentiation of human genome-scale metabolic reconstructions [42, 43] into specific tissue types, or the approaches to develop microbial genome-scale metabolic models for individual species and communities [44].

During the construction of this model, we have especially focused on photosynthesis, since this is arguably the most important biological process occurring in nature, and also a process that is largely neglected, and sometimes even represented incorrectly in genome-scale metabolic reconstructions of photosynthetic organisms. Photosynthetic electron transport fuels the Calvin-Benson cycle, which is responsible for the ability of microalgae and higher plants to fix carbon in the form of CO2. This influx of carbon makes photosynthetic organisms able to produce large amounts of lipids, and thus makes them relevant targets for industrial lipid production. The chloroplast model can therefore be used to explore and compare modes of lipid production in the currently included microalgae modes, Nannochloropsis, Phaeodactylum and Chlamydomonas.

We have also developed software tools for incorporating the chloroplast model as a module into another metabolic model in need of a chloroplast, for changing the organism mode of the chloroplast model, and for further expanding the model, either by adding organism-specific reactions for one of the organisms already included, or by adding a new organism to the model.

Model structure.

XML file containing the iGR774 model.

(XML)

Click here for additional data file.

Modelling tools developed alongside with the iGR774 model.

Zip-file containing modelling tools developed for MATLAB. The individual tools are described in S1 Text.

(ZIP)

Click here for additional data file.

Descriptions of modelling tools.

Detailed description of the modelling tools included in S2 File.

(DOCX)

Click here for additional data file.

Composition of chloroplast biomass components.

Detailed composition of the protein, membrane lipids and pigment fractions included in the chloroplast biomass component.

(DOCX)

Click here for additional data file.

Restricted reaction.

List of reactions with restricted upper or lower bounds, including reason for the restriction.

(XLSX)

Click here for additional data file.

Critical genes.

List of genes that proved to be critical to normal growth of the iGR774 model.

(XLSX)

Click here for additional data file.

Photosynthesis and Calvin-Benson cycle fluxes.

List of fluxes resulting from optimization of a mini-model consisting of only the photosynthetic electron transport chain and the Calvin-Benson cycle.

(XLSX)

Click here for additional data file.

9 Dec 2019

PONE-D-19-26969

An adjustable algal chloroplast plug-and-play model for genome-scale metabolic models

PLOS ONE

Dear Dr. Bjerkelund Røkke,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Jan 23 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Andrew Webber

Academic Editor

PLOS ONE

Journal Requirements:

1) When submitting your revision, we need you to address these additional requirements.Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2) PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

3)  Thank you for stating the following financial disclosure:

No. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

a) Please provide an amended Funding Statement that declares *all* the funding or sources of support received during this specific study (whether external or internal to your organization) as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now.

b) Please state what role the funders took in the study.  If any authors received a salary from any of your funders, please state which authors and which funder. If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

Additional Editor Comments:

The manuscript needs some very significant revisions as suggested by the comments below. Hopefully, these comments/suggestions can be used to constructively update the manuscript. It would be very valuable to include import of nuclear encoded components into the model.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript as written provides little new insight into the metabolic fluxes/metabolism taking place within chloroplasts of algae. A lot of statements are obvious or text book level. For example line 359-372, 410-431, 489-493 are all basic knowledge of photosynthesis.

The authors should add import of substrates from the cytosol to the model. This is how metabolism actually takes place within algae. This also would allow simulation of dark conditions in which an external or storage substrate is utilized.

The biomass equation for the chloroplast appears to be missing critical components. Where is the DNA/RNA present in the plastid? Do the authors allow for chloroplast proteins to be important from the cytosol? A major fraction of the biomass of chloroplasts comes from nuclear encoded proteins.

The authors should be more certain about the result presented in line 345. Can't one remove the pigment synthesis unique to the P and C models and see if the growth rates become equal?

What do you mean by line 378 "at best"?

What is incorrect about models regarding proton pumping (line 384)? FBA doesn't need concentrations, only that the stoichiometry is correct.

Figure 2 is too difficult to read.

Figure 3 and 4 could be represented by just 2 numbers representing the slopes. There are no shifts in slope over different photon levels. In fact Figure 4 seems redundant with 3.

How do the results in Fig. 3 and 4 compare to photons/chloroplast? chlorophylls/chloroplast? In other words, how much light is absorbed by each g DW? Is it the same for each organism? Line 408 reads poorly. Photons aren't generated

The line 455 where the model is forced to produce NADPH and ATP in a specific ratio makes the result on line 458-459 meaningless. Of course the ration needed by the CBB is met. Could the authors comment for splitting 1 H20 how many NADPH are made and how many protons are released and what stoichiometry of the ATP synthase of protons per ATP is used.

Figure 5 the y-axis is a growth rate, but the text (line 500) says it should be lipid production.

Most of the conclusions are not conclusions; e.g 518-520, 520-322 are basic statements true without the work in the manuscript. The rest of the conclusions are just a restatement or summary of result.

A minor recommendation (not required) on the tags that would make the model more generalizable is that "C" is not specific enough as many algae start with C, like Chlorella. A three or four letter abbreviation would be more appropriate.

Is the model capable of simulating chromoplasts as well as chloroplasts?

Don't start sentences with #. e.g. line 321

Line 486 please check the wording-"combined pool important chloroplast-produced" the word important doesn't makes sense. Do you just mean total lipids?

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: John A Morgan

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

8 Jan 2020

We would like to thank Dr. Morgan for his comments and suggestions, and for taking the time and effort to review our manuscript. We have responded to his comments below (responses are marked >>>), and we hope that our revised manuscript will be considered suitable for publication.

Comments from Reviewer:

The manuscript as written provides little new insight into the metabolic fluxes/metabolism taking place within chloroplasts of algae. A lot of statements are obvious or text book level. For example line 359-372, 410-431, 489-493 are all basic knowledge of photosynthesis.

>>> We apologize for leaving the reader with the expectation that novel insights should be gained in this section. Indeed, the textbook level statements are intended to describe the detailed, known processes that are in fact included in our model. A challenge with previous models has been that they have been lacking in this respect. Our primary focus has been to develop a tool that will aid model building of organisms that contain chloroplasts, and to offer a new approach to inclusion of organelles into an existing model. We indeed emphasize in the manuscript that this model is a proof of principle, offering a new approach to model building, and that it is meant to be expanded by the future users of our model module. For this purpose, we included tools for expanding the model with new reactions in the toolbox found in S2 File (addRxFromModelSyntaxGenerator.m, pickKEGGrx.m and rxGenerator.m).

Consequently, including basic knowledge of photosynthesis in the manuscript is deliberate on our part, since we assume that the main interest in our manuscript will come from scientists working with modelling, and not necessarily from plant biologists. We also put a special emphasis on photosynthesis because we see that the photosynthetic processes are often incorrectly modelled in genome-scale metabolic reconstruction of photosynthetic organisms. It thus seems that even though this knowledge is on a textbook level, it has not found its way from the textbooks and into the published genome-scale reconstructions. We have now made our intentions clearer by including a statement about our intent on lines 86 – 93 and 95 – 102.

The authors should add import of substrates from the cytosol to the model. This is how metabolism actually takes place within algae. This also would allow simulation of dark conditions in which an external or storage substrate is utilized.

>>> We very much agree that import from cytosol to the chloroplast module is important. In fact, most of the substrates used by the chloroplast model are imported from cytosol. In order to differentiate between metabolites belonging to the cytosol and those of the chloroplast module, our model thus also contains a cytosol version of the metabolites that are either imported or exported from the chloroplast. When the chloroplast module is plugged into an exo-model, the script being responsible for joining the models (plugAndPlay.m) searches for these metabolites in the cytosol of the exo-model, and translates the metabolites exported from and/or imported to the chloroplast model into the namespace of the exo-model. This is done to ensure that the transport metabolites are indeed able to be transported between the cytosol of the exo-model and the chloroplast model, and to ensure that the exchange between the chloroplast module and the exo-model is transparent. We also consider this approach very helpful for debugging purposes.

The only 'metabolites' that are not imported from cytosol are the photons used by in the photosynthetic reactions, since light is not a conventional metabolite being transported between organelles. Importing photons directly to the chloroplast also ensures that the combined model consisting of the exo-model and the chloroplast model will run, even if the exo-model does not import photons. We have now added a statement to clarify this issue on lines 162 – 172.

The biomass equation for the chloroplast appears to be missing critical components. Where is the DNA/RNA present in the plastid? Do the authors allow for chloroplast proteins to be important from the cytosol? A major fraction of the biomass of chloroplasts comes from nuclear encoded proteins.

>>> We thank the Reviewer for bringing up this observation. Synthesis of RNA and DNA has been omitted in the current version of the model because these reactions are not essential to chloroplast function. Instead, we have focused on the synthesis of lipids, pigments and proteins, since these are metabolites that are currently attracting a lot of scientific interest. Certain fatty acids that are only synthesized in the chloroplast are transported to other parts of the cell and used further, and the chloroplast also synthesizes several essential amino acids. The generated lipids and proteins / amino acids are therefore metabolite groups that the chloroplast needs to produce in order for the combined model to run, while the same cannot be said about RNA and DNA.

The model in its current form is intended as a proof of principle. Specific reactions related to production of RNA and DNA that would be of interest for e.g. a particular exo-model can of course be added via the tools we provide for adding new reactions to the model. We have now added a statement to explicitly address this point on lines 134 – 140.

The authors should be more certain about the result presented in line 345. Can't one remove the pigment synthesis unique to the P and C models and see if the growth rates become equal?

>>> We thank the Reviewer for bringing up this interesting point. Even if the organism-specific pigment component was removed from the chloroplast biomass reaction, the Phaeodactylum-specific and the Chlamydomonas-specific growth rates would still not be equal. One of the reasons for this is that Chlamydomonas has an active eyespot, and the metabolism of the eyespot is included in the Chlamydomonas mode, but not in the Phaeodactylum mode. In addition, there are differences in the composition of proteins and lipids. Proteins in Phaeodactylum and Chlamydomonas contain the same 20 amino acids, but in different ratios. The ratio between different fatty acids in the Phaeodactylum and Chlamydomonas specific pools of MGDG, DGDG and SQDG is also different. Both Phaeodactylum and Chlamydomonas are able to produce C22:6 (Nannochloropsis is not), but Chlamydomonas does for example produce C18:3 (now stated in the manuscript, at lines 376 – 378), which is not present in Phaeodactylum (if it is present, it is assumed to be so at negligible amounts).

Thus, even if pigment synthesis would be removed, the growth rates of Phaeodactylum and Chlamydomonas would therefore still not be equal, as the difference in the ratios of amino acids in proteins, and of fatty acids in different lipid pools will create a shift in the fluxes of individual reactions that would be obvious if the flux vectors resulting from optimizations of the two organisms were compared.

We also state in the manuscript that the composition of lipids and proteins are dissimilar in the different organism modes (lines 378 – 383), and we refer to the exact composition in S4 Text (line 383).

What do you mean by line 378 "at best"?

>>> We thank the Reviewer for bringing up this unfortunate choice of phrase. We mean that in most published models of photosynthetic organisms, photosynthesis is mostly described by less than ten reactions. We have now changed this phrase in the manuscript to hopefully make our intention clearer (line 431).

What is incorrect about models regarding proton pumping (line 384)? FBA doesn't need concentrations, only that the stoichiometry is correct.

>>> We very much agree with the Reviewer's observation of FBA properties. Here, if the number of protons pumped across the thylakoid membrane is not correct, this does affect the stoichiometry, and in particular the ratio between ATP and NADPH produced during photosynthesis. In several published models of photosynthetic organisms, the ratio between ATP and NADPH produced in photosynthesis is incorrect, and the reason for this is that the net number of protons transported across the thylakoid membrane during the production of one molecule of NADPH is not correct.

Figure 2 is too difficult to read.

>>> We thank the Reviewer for bringing up this point. We have now tried to make the figure more readable: The font size has been increased, and the text has been made bold. In addition, nodes representing ‘metabolites’ and ‘reactions’ have gotten separate colours. We have also marked the reactions taking place in the different photosynthetic complexes. The figure text on lines 529 – 533 has been changed to properly explain the new version of the figure.

Figure 3 and 4 could be represented by just 2 numbers representing the slopes. There are no shifts in slope over different photon levels. In fact Figure 4 seems redundant with 3.

>>> Here, we are of the opinion that a figure may be a more comprehensible way of illustrating a phenotype phase plane, instead of just drawing two slopes. Regarding redundancy of figures 3 and 4; we agree with the Reviewer’s point. We have chosen to only keep the figure labelled figure 4 in the first submission of the manuscript, showing growth rate as a function of photon usage by PSII and the rate of carbon fixation by the Calvin-Benson cycle. This figure has now been renamed Fig 3 (lines 458 and 535), and the old figure 5 has been renamed figure 4 (lines 570, 574 and 581).

How do the results in Fig. 3 and 4 compare to photons/chloroplast? chlorophylls/chloroplast? In other words, how much light is absorbed by each g DW? Is it the same for each organism?

>>> In both Nannochloropsis, Phaeodactylum and Chlamydomonas mode, photosynthesis is running at the maximum allowed flux for photon import, meaning that 1000 mmol of photons is imported to the chloroplast per gram dry weight per hour. We have now included a statement to clarify this point on lines 467 – 470.

Line 408 reads poorly. Photons aren't generated

>>> We thank the Reviewer for catching this unfortunate statement. We have now corrected this sentence in the revised version of the manuscript (lines 465 – 466).

The line 455 where the model is forced to produce NADPH and ATP in a specific ratio makes the result on line 458-459 meaningless. Of course the ration needed by the CBB is met. Could the authors comment for splitting 1 H20 how many NADPH are made and how many protons are released and what stoichiometry of the ATP synthase of protons per ATP is used.

>>> We thank the Reviewer for bringing up this point. The reason for restricting the rate of cyclic electron transport was that the photosynthetic processes did not produce NADPH and ATP in the ratio that the Calvin-Benson cycle needed -- photosynthesis did, in fact, not produce NADPH at all. This is why we saw that it might be necessary to restrict the rate of cyclic electron transport, and we built the mini-model consisting only of the photosynthetic electron transport chain and the Calvin-Benson cycle in order to find the most sensible upper bound for the reaction being responsible for cyclic electron transport.

1 molecule of NADPH is created per H2O molecule splitted. Simultaneously, 12 protons are pumped into the thylakoid lumen during one complete S-cycle, meaning that the H+:H2O ratio is 6:1. Regarding ATP synthase, most ATP synthases require 12 protons to be pumped from the lumen and back to the stroma in order for its γ subunit to make a 360 degree movement, which will regenerate 3 molecules of ATP, yielding an ATP:H+ ratio of 1:4. These ratios can also be calculated from table S7. In addition, we state in the manuscript that when cyclic electron transport is eliminated, the photosynthetic electron transport chain produce the exact ratio of ATP and NADPH that the Calvin-Benson cycle needs (lines 521 – 523), and on lines 475 – 477 we state that the ATP:NADPH ratio needed by the Calvin-Benson cycle is 9:6.

Figure 5 the y-axis is a growth rate, but the text (line 500) says it should be lipid production.

>>> We thank the Reviewer for catching this mistake. We have now changed the y-axis label in the figure (now figure 4) to 'rate of lipid production'.

Most of the conclusions are not conclusions; e.g 518-520, 520-322 are basic statements true without the work in the manuscript. The rest of the conclusions are just a restatement or summary of result.

>>> We apologize for this, and we have now expanded the conclusion (lines 587 – 603, 609, 615, 617 – 620 and 784 - 794). We have now tried to emphasize what is new about our model, and what the advantages of organelle modules such as our chloroplast module are. We have also included a few extra references, to add some context, and to give an example of other approaches being used by the modelling community to represent differences in metabolism between cells.

A minor recommendation (not required) on the tags that would make the model more generalizable is that "C" is not specific enough as many algae start with C, like Chlorella. A three or four letter abbreviation would be more appropriate.

>>> We agree with the Reviewer, especially since several of the tools we have developed are meant to facilitate further expansion of the model by adding new reactions and new organism modes. We have now changed the tags of the model into Nan for Nannochloropsis, Pha for Phaeodactylum, and Chl for Chlamydomonas. The organism tags has been changed both in the reaction vector of the model, in the four scripts using organism tags (addRxFromModelSyntaxGenerator.m, changeOrganismMode.m, pickKEGGrx.m, and rxGenerator.m (All scripts are found in S2_File), and in the explanation of the organism tags in the manuscript (lines 121 – 123).

Is the model capable of simulating chromoplasts as well as chloroplasts?

>>> To our knowledge, chromoplasts are mainly associated with storage of pigments, while photosynthesis does not play a big part. Especially in the chromoplasts found in plant roots, we assume that photosynthesis is absent. We put a lot of emphasis on the photosynthetic electron transfer chain during model development. Yet, if the chloroplast model is plugged into an exo-model that does not require it to produce ATP and NADPH photosynthetically, we see no reason why a chromoplast organism mode could not be introduced to the chloroplast model. Synthesis reactions of several pigments are already present in the chloroplast module, and other synthesis pathways could be introduced in a new organism mode. For getting the chloroplast module to store metabolites, a common solution would be to introduce a direct “exchange” reaction (transport out of the cell) for each of the metabolites to be stored – thus not breaking the steady-state assumption.

Don't start sentences with #. e.g. line 321

>>> We thank the Reviewer for this comment, and have now changed the beginning of these sentences in the revised manuscript (lines 360, 364 – 365, 367, 369 – 370 and 372).

Line 486 please check the wording-"combined pool important chloroplast-produced" the word important doesn't makes sense. Do you just mean total lipids?

>>> We thank the Reviewer for catching this mistake. It seems the sentence was missing a word. This has now been corrected (line 561).

Submitted filename: Rebuttal_letter.docx

Click here for additional data file.

6 Feb 2020

An adjustable algal chloroplast plug-and-play model for genome-scale metabolic models

PONE-D-19-26969R1

Dear Dr. Bjerkelund Røkke,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Andrew Webber

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for your careful set of responses.

Reviewers' comments:

10 Feb 2020

PONE-D-19-26969R1

An adjustable algal chloroplast plug-and-play model for genome-scale metabolic models

Dear Dr. Bjerkelund Røkke:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Andrew Webber

Academic Editor

PLOS ONE