As a basis for the in silico rice modeling project, e-Rice, the preliminary goal is to develop a model and simulate plant cell primary metabolism using E-CELL simulation environment. In order to accomplish modeling of such large scale, a basic framework for modeling must be made. A model of the most basic plant cell metabolism has been constructed using available databases and automated literature search system. A more comprehensive model for plant cell primary and secondary metabolism is currently being constructed based on genome and metabolome data. A detailed model of the photosynthetic electron transport chains is also being constructed based on literature.

As a basis for the e-Rice in silico rice modeling project, the preliminary goal is to develop a model for simulating plant cell primary metabolism using E-CELL simulation environment. In order to accomplish modeling of such large scale, a basic framework for modeling must be made, whereby the modeling level is determined based on several factors, including data availability, level of abstraction and scope for data usage. Modeling a eukaryotic cell pose numerous obstacles due to its diversity and complexity. Cellular structure varies in relation to function and location, with metabolism dependent on hormonal elements such as growth phase and environmental effects. Though abundant data exist on plants, only a limited number is relevant to the goal of the project. One priority is to devise methods which can maximize the yield of information. Our first assignment in undertaking the enormous task is to grasp the components involved in eukaryotic cellular modeling, without which, efficient modeling of a cohesive system would be impossible.

Modeling the plant cell poses numerous challanges. These diffulties can be attributed to the location of the plant cell, which is heavily intertwined with its structure and function. Plants also show continuous changes in gene expression and enzyme activity depending on environmental conditions. As higher plants exhibit such continuous changes, steady-state conditions often do not apply when modeling cellular conditions (Morgan.et.al, 2002).

1. Cell Structure, Function and Location: Cell structure varies with its required functions. These functions are facilitated by its location.
2. Mesophyll cell: located on the upper surface of the leaf. Its location allows efficient absorption of light, containing the most amount of chloroplasts compared to other cell types (Lawlor, 1993).
3. Guard cell: The guard cell function is to control and stabilize H2O evaporation from the leaf. Such mechanisms are regulated by ionic channels, that are well studied.
4. Parenchyma cell: Located next to the phloem, parenchyma cells are central in sucrose transport from other surrounding cells of the leaf.
5. Bundle-sheath cell: Constituting about 15% of chloroplast containing cells, bundle-sheath cells are adjacent to the vein with different plastid development from mesophyll cells (Kinsman.et.al, 1998).

The plant cell constitutes of various organelles under the family of plastids; these include the chloroplast, amyloplast, etioplast and others. Much like stem cell development, these plastids are the developed form of a proplastid, each resulting from different growth conditions and having different functions. While mesophyll cells develop photosynthetic chloroplasts in the presence of light, etioplasts are developed under dark conditions. Root cells contain amyloplasts, a starch accumulating plastid. Development of plastids are controlled by plant hormones, with cytokinin working as a positive regulator for chloroplast development(Nakano.et.al, 2001). Representation of the complete secondary metabolism may be difficult due to limited data. However, it is also essential to model certain secondary metabolisms without with the plant cell cannot function.

The attempt to accumulate all pathway data from literature, though efficient in analyzing and grasping the kind of data available for each pathway, is far from enough at comprehensive data collection.

DBGET at KEGG using O.sativa database, release 22.0+/06-10, Jun 00, contains 225 entries for rice, while it contains a significantly larger amount for arabidopsis. Though synteny between Arabidopsis and rice genes were limited, 2565 predicted protein genes showed high conservation between Arabidopsis and rice genes. In addition Arabidopsis is known to lack certain gene classes that are found in rice(Goff.et.al, 2002). The combined use of rice and arabidopsis pathways from KEGG would be reasonable within accepatble scale for pathway assumptions until detailed information from rice cDNA becomes available.

Automated pathway retrieval and model construction is currently being developed in conjunction with the Sequence Analysis Group at the Bioinformatics Program, and the E-CELL E2coli project.

As a result of sequence homology, Goff.et.al noted that approximately 25% of genes in rice are involved in metabolism. At this point, genomic information is one of the key method in accumulating all relevant protein function in comparison with other annotated genes. Various databases and softwares geared to the rice genome are readily availble and will be used in conjunction with methods above to provide a data map for all significant metabolic proteins, as well as signal transduction, ion transporters and other related regulators that have been identified in the rice genome. A list of such sites is listed below.

• RiceGAAS: Rice Genome Automated Annotation System - http://ricegaas.dna.affrc.go.jp/
• The TIGR Rice Genome Project - http://www.tigr.org/tdb/e2k1/osa1/
• Monsanto rice-research - http://www.rice-research.org/
• The Korea Rice Genome Database - http://bioserver.myongji.ac.kr/ricemac.html
• Oryzabase - http://www.shigen.nig.ac.jp/rice/oryzabase/
• Rice cDNA Sequence Analysis Project - http://www.cbc.umn.edu/ResearchProjects/Rice/index.html
• US Rice Genome Sequencing - http://www.usricegenome.org/
• Beijing Genomics Institute Rice Genome - http://btn.genomics.org.cn/rice/index.php
• HarvEST - http://harvest.ucr.edu/

Using an automated literature retrieval and selection system developed by an undergraduate student Kotaro Ishii , a preliminary search was conducted on rice metabolism. Search conditions were loosely set, with literatures at all growth phases and locations included. However, the search was restricted only to rice and other cereals. Initial search research results were manually screened to eliminate irrelevant articles. Keywords that are prevalent in irrelevant articles were used as feedbacks to the literature search system to prevent future retrieval of such articles.

The results were not as distributed and comprehensive as expected, in terms of modeling-geared data. Kinetic and time-series data required for modeling are mostly centered on energy metabolism and a few on amino acid metabolism. Data on other metabolisms are scarce, with an exception of certain enzymes and pathways that are well studied, such as alpha-amylase in rice seeds, primary metabolism and photosynthesis.

The draft plant energy metabolism model (e-Rice Ver.0) consists of metabolisms in the chloroplast, cytoplasm and the mitochondrion, with a total number of 141 reactions including the following pathways and compartments:

• Cytoplasm:
  • Hexose Mono-Phosphate (HMP) shunt
  • Embden-Meyerhof-Parnas (EMP) pathway
  • Sucrose synthesis
• Mitochondria:
  • TCA Cycle
  • Beta-oxidation
  • Transport carriers
• Chloroplast:
  • Calvin Benson cycle of photosynthesis
  • Starch synthesis

A new metabolism model is currently being modeled from various literature(Lawlor.et.al, Rice.et.al, Foyer.et.al). Reactions involve a total of six compartments; cytoplasm, chloroplast, mitochondrion, peroxisome, spherosome, and vacuole. Though the goal is to cover primary metabolism, several secondary metabolisms are essential for plant survival. These essential metabolisms such as cytokinin will also be included when identified.

Photosynthesis, a process which combines biological, physical and chemical properties to assimilate and convert energy from sunlight to biochemical energy, is one of the most intriguing phenomenon found in organisms. Though some aspects of photosynthesis are well studied, other components are not well known and have yet to be classified. Recent advances in technology have greatly aided research in this area, enabling the characterization of certain key features in photosynthesis (Baker.et.al.2001, Gomez.et.al.2002). Furthermore, these features are not only essential in converting light energy, but their regulatory functions have been found to greatly affect other elements in cellular systems such as metabolism, signalling, and gene expression (Pfannschmidt.et.al.2001, Trebitsh.et.al.2001, Sherameti.et.al.2002, Karpinski.et.al.1999). Creating a model that takes into account these regulatory elements is essential in further understanding of photochemical processesand how they affect the chloroplast, the plant cell, and the whole plant.

The entire process of photosynthesis occurs in two phases, the light reactions and the dark reactions, also known as the Calvin cycle. The light reactions take place on the thylakoid membrane, assimilating light energy to generate O2, NADPH and ATP through an electron transport chain. These products are used in the dark reactions and other carbon metabolisms, incorporating carbon dioxide to yield products such as starch, sucrose, and other metabolites. Using the E-CELL system, a prototype model of the Calvin-Benson photosynthesis cycle has previously been constructed based on mathematical models of Pettersson et al (1989) and Laisk et al (1989). The model contains 39 metabolites and 32 reactions, 10 of which are characterized as equilibrium. The model also includes pathways for starch synthesis in the stroma, sucrose synthesis in the cytoplasm and also antiport transport for triose phosphates from the stroma to the cytoplasm.

Though the dark reactions have been characterized and parametrized in various models (Laisk.et.al.1988, Pettersson & Ryde-Pettersson.1988, Hahn.et.al.1986, Woodrow(1986), Giersch.et.al.(1991)), most models either abstract, or exclude the light reactions as a separate system of reactions. Following some enlightenment regarding membrane proteins on the thylakoid, the elements of the light reaction electron transport chain, it has also been found that the light and dark reactions are heavily intertwined and co-regulated. The modeling of dark reactions alone is insufficient in expressing the mechanisms of photosynthesis as a whole.

We are currently investigating the following models, which express photosynthetic light reactions, including those represented as abstract elementrary models or as detailed electron transport models. Their representation, efficacy and reliability are considered as indices of model quality.

• Farquhar, von Caemmerer & Berry (1980)
• Gutschick (1984)
• Sage (1990)
• Gross, Kirschbaum & Pearcy (1991)
• Fridyland & Schiebe (1999)
• Berry & Rumberg (2000)

A general framework for developing a plant cell metabolism has been created, with a large amount of information retrieved from literature and databases using automated systems. These data have been used to construct a draft model of primary plant cell metabolism. As rice-specific literature is extremely limited, data from various plants are currently being used for modeling.

Methods and tools to automate E-CELL modeling based on the rice genome and various other databases are also being developed for top-down modeling. Together with detailed modeling of individual pathways based on literature, and in collaboration with the metabolome experiment team at the Institute for Advanced Biosciences, a bottom up approach to modeling is being done concurrently.

The development of other tools and methods will be necessary, including:

• Data management system
• Methods on verifying the model
• Analysis tools
• Applications of crop modeling at a molecular level

The detailed model of photosynthesis light reactions is currently being constructed, representing major photosynthetic elements at a molecular level rather than the abstrated form of electron transport products. The model will be integrated into the next version of e-Rice model, together with the genomics derived top-down model integrated with data from the metabolomics team.