Waddington redux: models and explanation in stem cell and systems biology

Stem cell biology and systems biology are two prominent new approaches to studying cell development. In stem cell biology, the predominant method is experimental manipulation of concrete cells and tissues. Systems biology, in contrast, emphasizes mathematical modeling of cellular systems. For scientists and philosophers interested in development, an important question arises: how should the two approaches relate? This essay proposes an answer, using the model of Waddington’s landscape to triangulate between stem cell and systems approaches. This simple abstract model represents development as an undulating surface of hills and valleys. Originally constructed by C. H. Waddington to visually explicate an integrated theory of genetics, development and evolution, the landscape model can play an updated unificatory role. I examine this model’s structure, representational assumptions, and uses in all three contexts, and argue that explanations of cell development require both mathematical models and concrete experiments. On this view, the two approaches are interdependent, with mathematical models playing a crucial but circumscribed role in explanations of cell development.     

Generation of Mouse Small Intestinal Epithelial Cell Lines That Allow the Analysis of Specific Innate Immune Functions

Cell lines derived from the small intestine that reflect authentic properties of the originating intestinal epithelium are of high value for studies on mucosal immunology and host microbial homeostasis. A novel immortalization procedure was applied to generate continuously proliferating cell lines from murine E19 embryonic small intestinal tissue. The obtained cell lines form a tight and polarized epithelial cell layer, display characteristic tight junction, microvilli and surface protein expression and generate increasing transepithelial electrical resistance during in vitro culture. Significant up-regulation of Cxcl2 and Cxcl5 chemokine expression upon exposure to defined microbial innate immune stimuli and endogenous cytokines is observed. Cell lines were also generated from a transgenic interferon reporter (Mx2-Luciferase) mouse, allowing reporter technology-based quantification of the cellular response to type I and III interferon. Thus, the newly created cell lines mimic properties of the natural epithelium and can be used for diverse studies including testing of the absorption of drug candidates. The reproducibility of the method to create such cell lines from wild type and transgenic mice provides a new tool to study molecular and cellular processes of the epithelial barrier.     

Infrared studies of fully hydrated unsaturated phosphatidylserine bilayers. Effect of Li+ and Ca2+

Infrared spectroscopy has been used to characterize the thermal-phase behavior of fully hydrated 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS) and 1,2-dioleoyl-sn-glycero-3-phospho-L-serine (DOPS) as well as their interaction with Li+ and Ca2+. The order-disorder transition of POPS-NH4+ is at 17 degrees C; in the presence of Li+ a POPS-Li+ complex is formed, and the transition temperature of this complex is 40 degrees C. DOPS-NH4+ has an order-disorder transition at -11 degrees C, and unlike POPS the addition of Li+ has no effect on the thermal behavior of DOPS-NH4+. This indicates that the binding of Li+ to DOPS is negligible or very weak. Li+ binds to the phosphate and carboxylate groups of POPS, and as a result these groups lose their water of hydration. Li+ binding induces a conformational change, probably in the glycerol backbone of POPS; however, the conformation of the two P-O ester bonds remains gauche-gauche as in POPS-NH4+. Both POPS and DOPS form crystalline complexes with Ca2+. As a result of Ca2+ binding to the phosphate, this group loses its water of hydration and there is a conformational change in the P-O ester bonds from gauche-gauche to antiplanar-antiplanar. In contrast to the POPS-Li+ complex, the carboxylate group remains hydrated in the Ca2+ complexes. Furthermore, in these PS-Ca2+ complexes a new hydrogen bond is formed between one of the ester C=O groups and probably water. Such a situation is not found in the NH4+ and Li+ salts of phosphatidylserine.     

Studies on direct and indirect effects of DNA damage on mutagenesis in monkey cells using an SV40-based shuttle vector

We are using an SV40-based shuttle vector, pZ189, to study mechanisms of mutagenesis in mammalian cells. The vector can be treated with mutagens in vitro and replicated in animal cells; resulting mutants can be selected and amplified in bacteria for DNA sequencing. This versatile vector system has allowed us to explore several different questions relating to the mutagenic process. We have studied the direct effects of template damage caused by UV or benzo[a]pyrene diolepoxide by treating vector DNA with these agents and then replicating the damaged DNA in monkey cells. Mutational mechanisms were deduced from the spectrum of mutations induced in the supF target gene of the vector DNA. To study the role of indirect effects of DNA damage on mutagenesis in mammalian cells, we have treated the cells and the vector DNA separately with DNA-damaging agents. We find that pretreatment of cells with DNA-damaging agents, or with conditioned medium from damaged cells, causes an enhancement of mutagenesis of a UV-damaged vector. Thus, DNA damage can act indirectly to enhance the mutagenic process. We also have preliminary evidence that pZ189 can be used in an in vitro DNA replication system to study the process of mutation fixation on the biochemical level. We believe that the pZ189 vector will prove to be as useful for in vitro studies of mutational mechanisms as it has been for in vivo studies.     

Interactions of metal ions with phosphatidylserine bilayer membranes: effect of hydrocarbon chain unsaturation

A combination of surface monolayer, scanning calorimetry, 31P NMR, and spin-label ESR techniques has been used to monitor the interactions of monovalent (NH4+, Na+, and Li+) and divalent (Ca2+) cations with phosphatidylserines (PS) differing in their levels of chain unsaturation. Comparisons are made between the disaturated dimyristoyl-, dipalmitoyl-, and dihexadecyl-PS (DMPS, DPPS, and DHPS), saturated cis-monounsaturated palmitoyloleoyl-PS (POPS) (and bovine brain PS), di-trans-monounsaturated dielaidoyl-PS (DEPS), and di-cis-monounsaturated dioleoyl-PS (DOPS). Na+ and NH4+ cations interact weakly with all PS monolayers and bilayers without significant changes in molecular conformation, chain packing, or headgroup dynamics and without dependence on chain composition. In contrast, considering these structural and dynamic parameters, Li+ shows a gradation in its interaction with PS (DMPS greater than POPS approximately bovine brain PS greater than DOPS), suggesting that Li+-PS interactions depend on the interfacial properties of the PS molecules (e.g., surface area). Finally, Ca2+ interacts strongly with all PS monolayers and bilayers, without obvious chain selectivity. Thus, ion binding to PS depends not only on the properties of the cation (Na+ vs Li+ vs Ca2+) but also on the molecular details of the PS membrane surface.     

Thiophosphorylated proteins in chromaffin cells are chromaffin vesicle matrix proteins

Bovine adrenal chromaffin cells were incubated with inorganic thiophosphate, using a protocol similar to experiments with inorganic phosphate, in order to determine the source of previously observed thiophosphoproteins. Incubation of cultured cells with [³⁵S]thiophosphate resulted in its incorporation into cell constituents within 2 min. SDS PAGE of the treated cells showed incorporation of label into a broad 97–121 kDa band that was evident after 5 min of treatment and increased progressively to the 40 min exposure limit. Monolayers of chronically treated cells were fractionated into subcellular constituents. The only particulate fraction containing radiolabelled proteins was the chromaffin vesicle fraction. Two-dimensional electrophoresis of the treated cells and isolated chromaffin vesicles showed a majority of proteins in the acidic region of the first dimension gel. A fluorogram of the gel revealed two regions of radiolabelled proteins at acidic and neutral regions of the 2-D gel. These were within the boundaries of the 97–121 kDa band. The thiophosphorylated proteins were released as soluble proteins upon osmotic or freeze-thaw lysis of the vesicles. Chromaffin vesicles isolated from either cultured cells or adrenal medulla tissue were energized by 2 mM ATP but not by the analog adenosine 5′-O-(3-thiotriphosphate). The 97–121 kDa proteins in intact or lysed vesicles prepared from adrenal medulla tissue were not thiophosphorylated by either inorganic thiophosphate or adenosine 5′-O-(3-thiotriphosphate) in the presence or absence of energization by ATP. Nearly complete loss of radiolabel from matrix proteins treated with chondroitinase ABC suggests that it is a component of vesicle proteoglycans.

The results demonstrate that chromaffin vesicle matrix proteins are rapidly and intensely thiophosphorylated in cultured chromaffin cells but not in isolated vesicles. The data suggest that phosphorylation must play an important role in the normal function of these vesicle proteins.     

Aromaticity at position 37 in human epidermal growth factor is not obligatory for activity.

The third disulfide loop (amino acids 33 to 42) of human epidermal growth factor (hEGF) encompasses the region of highest amino acid conservation among all of the EGF-like family of molecules. The importance of some of these highly conserved residues for the maintenance of biological activity, especially the aromatic amino acid tyrosine at position 37, has until now been considered essential on the basis of previous studies with the EGF-like molecule transforming growth factor alpha. Variants at the Tyr-37 position of hEGF were constructed by site-directed mutagenesis. The substituting amino acids were phenylalanine, histidine, serine, alanine, aspartic acid, arginine, and glycine. The variants were tested for their ability to competitively displace native [125I]hEGF from its receptor and to stimulate the protein-tyrosine kinase activity of the receptor; the order of efficacy of substituting amino acids was Phe greater than His greater than Ser greater than Ala greater than Asp greater than Arg greater than Gly in both assays. All were effective, with no or only moderate reduction in potency, in stimulating the incorporation of [3H]thymidine into acid-insoluble material of quiescent mouse A31 cells. Only Tyr-37----Ala, Tyr-37----Arg and Tyr-37----Gly were slightly less potent in the cell assay. Thus, neither tyrosine nor another aromatic amino acid at position 37 in hEGF is essential for full biological activity.