Ute Platzer et al. (2001) International C. elegans Meeting "Analyzing genetic networks that regulate the development of the early C. elegans embryo"

Ute Platzer, Hans-Peter Meinzer, Markus Gumbel

[International C. elegans Meeting, 2001]

The fate of a cell is determined by its "genetic state", by all proteins (or gene products) present in the cell. The genetic state of a cell can be represented by naming each gene and its current state, which can be either 1 ("on") or 0 ("off"). Cells change their states during development, according to the interactions between the genes. For example, end-1 is expressed as a result of the activity of the transcription factor SKN-1. A set of interactions is called a genetic network. Genetic networks can be modeled in (at least) two ways: First, as a matrix with as many rows and columns as there are genes in the network. The number in column i and row j represents the influence that gene no. i has on gene no. j . Second, as a boolean network: Each gene is assigned a logical operator, e.g. "AND" or "OR". Gene states serve as input values for other genes. We have developed a software system that allows the user to enter and analyze genetic networks. The analysis can be done using one of the methods mentioned above. The user can edit the network via a graphical interface. New genes can be added by simple mouse clicking, or they can be dragged with the mouse from a predefined "database"-network. Interactions between genes can also be added using the mouse. Alternatively, the user can directly specify interactions between genes by typing in the boolean function. The calculation of stable states of the network gives information about the number and kind of cell types that can be generated via this network. The user can thus check whether the genes he has entered so far are sufficient to obtain the different cell types observed in biological experiments. Furthermore, the user can enter data about the cells in the organism (e.g. radius and position), and he can define the initial states of all genes in every cell. Additional parameters, for example which gene is responsible for specifying the division ratio of a cell, can be fed in easily. The development of the worm embryo under control of the given network can then be simulated using the CellO program [1]. The cells reside within a 3-dimensional model of the egg shell. They divide according to the rules specified by the network. As the cells have a spatial relationship in this simulation, inductive interactions between cells are also possible. The user can define some genes in the network to be "external", meaning that they influence only genes in neighboring cells. It is also possible to set the cell positions as observed in wild type embryos [2]. This may be useful to study inductive interactions that depend on correct cell positions and spatial relations. To summarize, this piece of software can be used to (1) simply gather and sort information about genetic networks: a clear graphical view of the network is presented, (2) analyze networks: calculate stable states and watch cell and organism development, (3) make predictions about cell differentiation: by leaving out genes, the development of mutants may be predicted. We plan to implement a connection to internet databases, e.g. AceDB and WormBase, to import genetic data directly from the internet. As more and more data are collected, it may be possible to simulate the whole worm, not only the early embryo. References: [1] Gumbel M, Schnabel R, Meinzer HP (2000). 14th ESM Proceedings. SCS Publications, pp. 605-611. [2] Schnabel R, Hutter H, Moerman D, Schnabel H (1997). Dev Biol. 184(2), pp. 234-265.

Rolf Lohaus et al. (2003) International Worm Meeting "Simulation of the genetic network controlling C. elegans vulval development"

Hans-Peter Meinzer, Ricardo B R Azevedo, Rolf Lohaus, Ute Platzer

[International Worm Meeting, 2003]

Ricardo B R Azevedo et al. (2002) East Coast Worm Meeting "Nematode embryonic cell lineages are computationally efficient"

Markus Gumbel, Ricardo B R Azevedo, Ute Platzer, Hans-Peter Meinzer, Kim Jacobsen, Armand M Leroi, Volker Braun, Gaetan Borgonie, Paul-Michael Agapow, Wouter Houthoofd

[East Coast Worm Meeting, 2002]

What is the minimal amount of information required to specify the cells of a metazoan? Based on ideas from algorithmic information theory and phylogenetics, we develop an algorithm for predicting the distribution of determination events in complete cell lineages. We assume that all such events are either cell autonomous or the outcome of permissive cell-cell interactions, and that the lineage is parsimoniously specified. Applying our algorithm to the complete embryonic lineage of Caenorhabditis elegans, we show that it predicts many known molecular events required to specify cell fates. We then show that less information is required to specify the actual C. elegans lineage than lineages simulated under null models. This is also true for two other species of rhabditid nematode, Pellioditis marina and Rhabditophanes sp., despite many interspecific differences in lineage topology and cell fate assignments. Only one cell fate was found to be inneficiently specified in all species: programmed cell death. Unlike normal cells, most apoptotic cells appear to have no particular function during development. However, we show that the computational efficiency of embryonic development would be increased if cell deaths did not occur all. Thus, selection for increased computational efficiency should lead to a reduction in the number of programmed cell deaths in embryonic cell lineages. Although many programmed cell deaths occur in the C. elegans embryonic lineage (17% of all cells), all of them occur in single-cell monoclones. This is a significantly higher proportion than that expected from permuted lineages and suggests that cell deaths have not accumulated neutrally in the cell lineages of the ancestors of C. elegans. That the absence of cell death monoclones containing two or more cells is due to selection and not due to an intrinsic constraint is demonstrated by the observation that they have been found in other species. Such cases, we suggest, arise frequently, but are then eliminated by reprogramming. Indeed, the main function of somatic cell death in these nematodes might be to eliminate redundant cells over the course of evolution. Our results strongly suggest that selection for computational efficiency moulds the evolution of nematode embryonic cell lineages. But even though nematode lineages are more efficient than random lineages, they are clearly not as efficient as they might be. Why not? The polyclonal origin of some cell fates might be due to the need to generate cells of the same type, such as neurons, in various parts of the embryo. This is supported by the observation that in all species studied here, the majority of cells are born close to their final position in the embryo. We speculate that, in C. elegans, P. marina and Rhabditophanes sp., the fitness cost of repeatedly specifying the same cell type may be less than the cost of additional cell migrations.

Martin Gutternigg et al. (2006) European Worm Meeting "Glycosidases of Caenorhabditis elegans involved in N-glycan processing"

Martin Gutternigg, Matthias Hackl, Ute Stemmer, Iain B. H. Wilson, Petra Geier, Gunter Lochnit, Ramona Ranftl, Katharina Paschinger, Verena Jantsch, Dorothea Lubich

[European Worm Meeting, 2006]

Martin Gutternigg, Dorothea Lubich, Matthias Hackl, Katharina Paschinger, Ute Stemmer, Verena Jantsch1, Gnter Lochnit2, Ramona Ranftl, Petra Geier and Iain B. H. Wilson. Recent data indicates that in addition to the Golgi ?-mannosidases, the model nematode Caenorhabditis elegans also possesses, like insects, an N-acetylhexosaminidase activity putatively involved in N-glycan processing in the Golgi. The presence of such an activity is invoked, not just on the basis of the detected enzyme activity, but also to explain the absence of terminal N-acetylglucosamine residues on structures which require the prior action of N-acetylglucosaminyltransferase I during their biosynthesis. In order to understand the genetic basis for these activities, we have cloned cDNAs encoding members of both glycohydrolase families 20 and 38 from the worm. The encoded glycosidases were expressed in the yeast Pichia pastoris as soluble forms lacking putative cytoplasmic and transmembrane domains. Four glycohydrolase family 20 members were shown to cleave p-nitrophenyl-?-N-acetylglucosaminide and/or p-nitrophenyl-?-N-acetylgalactosaminide, but showed contrasting specificities with regard to N-glycan substrates. On the other hand, one glycohydrolase family 38 member was shown to be active using p-nitrophenyl-?-mannoside as a substrate and, in addition, had mannosidase II activity. Analysis of the glycans of the relevant mutant showed large-scale changes in the N-glycosylation spectrum. These, therefore, are the first data on the activity of Caenorhabditis glycosidases towards N-glycan substrates and should aid the further elucidation of N-glycan processing in this organism.