Conceptual modeling in systems biology fosters empirical findings: the mRNA lifecycle

Abstract

One of the main obstacles to understanding complex biological systems is the extent and rapid evolution of information, way beyond the capacity individuals to manage and comprehend. Current modeling approaches and tools lack adequate capacity to model concurrently structure and behavior of biological systems. Here we propose Object-Process Methodology (OPM), a holistic conceptual modeling paradigm, as a means to model both diagrammatically and textually biological systems formally and intuitively at any desired number of levels of detail. OPM combines objects, e.g., proteins, and processes, e.g., transcription, in a way that is simple and easily comprehensible to researchers and scholars. As a case in point, we modeled the yeast mRNA lifecycle. The mRNA lifecycle involves mRNA synthesis in the nucleus, mRNA transport to the cytoplasm, and its subsequent translation and degradation therein. Recent studies have identified specific cytoplasmic foci, termed processing bodies that contain large complexes of mRNAs and decay factors. Our OPM model of this cellular subsystem, presented here, led to the discovery of a new constituent of these complexes, the translation termination factor eRF3. Association of eRF3 with processing bodies is observed after a long-term starvation period. We suggest that OPM can eventually serve as a comprehensive evolvable model of the entire living cell system. The model would serve as a research and communication platform, highlighting unknown and uncertain aspects that can be addressed empirically and updated consequently while maintaining consistency.

Figure 1. A top-level view of the mRNA Lifecycle OPM model.

(A) An Object-Process Diagram (OPD) showing the top-level, bird's eye view of the process mRNA Lifecycle, which generates the object Protein from the object Amino Acid Set. The process, marked as the blue ellipse, shows the generation of mRNA from Ribonucleotide Set and the degradation of mRNA back to Ribonucleotide Set. (B) The corresponding Object-Process Language (OPL) text that was generated automatically by OPCAT, the software that supports OPM modeling.

Figure 2. Zooming into the mRNA Lifecycle process from Figure 1.

(A) An OPD in which the mRNA Lifecycle process from Figure 1 is in-zoomed into the three main subprocesses Transcription, Translation, and Post-ribosomal Processing. Time in an OPD flows from top to bottom, so Transcription is the initial process, Translation comes next, and finally Post-ribosomal Processing. While more detailed, the view of the system in this figure is fully consistent with its more abstract ancestor, shown in Figure 1. For the sake of simplicity transport-related processes are not depicted here yet. mRNA exhibits the attribute Localization, which can be at the states ribosome, nucleus, or P body. This relation between mRNA and its Localization attribute is denoted by the black-in-white triangle connecting these two objects. A thick process contour indicates that this process is in-zoomed in a different, more detailed OPD to show the inner content of that process. For example, the Translation process is zoomed in Figure 4. (B) The OPL text that explains the graphical model. For example, the OPL sentence Transcription occurs if Localization is nucleus. is denoted in the OPD by the link from the nucleus Localization to the Transcription process. In OPM, this link is called condition link, and it is denoted by the letter ‘c’ inside the link's circle end. In free text this means that transcription takes place in the nucleus.

Figure 3. Adding the mRNA transport processes.

(A) Three mRNA transport processes, marked in cyan, have been added to the OPD of Figure 2: Nucleo-cytoplasmic Transport, P Body-Ribosome Transport, and Ribosome-P body Transport. According to our model, the cytoplasmic mRNA is localized either in the ribosome or in the P bodies. For simplicity, possible transport from the nucleus to the ribosome is not modeled here. Therefore, the Nucleo-cytoplasmic Transport changes the Localization of the mRNA from nucleus to P body. The two other transport processes are the inverse of each other: P Body-Ribosome Transport, which changes the Localization of the mRNA from P body to ribosome, while Ribosome-P body Transport, which does the opposite. (B) The sentences of the OPL paragraph listing the processes and objects into which mRNA Lifecycle zooms and how each transport process changes the value of the Localization attribute of mRNA between nucleus, P body, and ribosome.

Figure 4. Zooming into the Translation process.

(A) Zooming into the Translation process exposes its three subprocesses that follow P Body-Ribosome Transport (initiation): Elongation, Termination, and Protein Cleavage&Releasing. The corresponding factors involved as instruments in these processes are also shown linked to the processes with an instrument link (a line ending with a circle at the process end). The object eEF Set is the instrument for the Elongation process, while eRF Set with its members, the factors eRF1 and eRF3, is the instrument for the Protein Cleavage&Releasing process. Our conjecture is that eRF3 is the factor which is also involved as instrument for the Ribosome-P body Transport process. Since there is no proof for this as yet, the instrument link from eRF3 to Ribosome-P body Transport is colored red, denoting uncertainty. (B) The OPL text of the OPD in (A). Note that the word requires in the OPL sentence Ribosome-P body Transport requires eRF3. denotes the same uncertainty regarding the role of eRF3 as instrument to the Ribosome-P body Transport process, analogous to the instrument link from eRF3 to Ribosome-P body Transport in (A).

Figure 5. Colocalization of eRF3-GFP with P bodies markers Dcp2p and Rpb4p.

Some P bodies are indicated by arrows (different kinds of arrows point at different P bodies). P bodies were examined after 7 days starvation in stationary phase as described in (Lotan et al., 2005). Merging of the green (eRF3-GFP) and the red (Dcp2p-RFP or Rpb4p-RFP, as indicated) channels was done using PhotoShop software. Note that if the green and red foci colocalize the resulting foci are yellow.

Figure 6. Zooming into Post-ribosomal Processing.

(A) Zooming into Post-ribosomal Processing exposes its subprocesses Checkpoint, Storing, and Degradation, as well as the objects D-factor and Decision. As before, red indicates uncertainty or hypothesis: We propose that D-factor is the instrument for the process we call Checkpoint, which in turn, determines whether to store, degrade, or unload the mRNA for reuse. Green links denote an uncertain conjecture that was confirmed in this work by our experiments. Here, the structural link from P Body to eRF3 and the tag contains along it are green, denoting that we demonstrated experimentally our model-based conjecture that P-body contains eRF3. (B) The OPL text of the OPD in (A). Note the red color of the words in the sentences Checkpoint requires D-factor. and in Checkpoint yields Decision. The red denotes that we are not certain whether Checkpoint and D-factor exist, and if so whether Checkpoint requires D-factor. On the other hand, the green color of the word contains in the sentence P-body contains eRF3 indicates our success at experimentally proving our model-based conjecture that P-body contains eRF3.

Figure 7. An example of modeling variants with OPM: A primary transcript is spliced to yield two splicing variants, A and B, each containing a different exon combination, as expressed in both the OPD at the top and the OPL text at the bottom.

Modeling other types of variants, e.g., protein phosphorylation, protein ubiquitination, RNA editing, can be done in a similar manner.