The role of body-on-a-chip devices in drug and toxicity studies.

High-quality, in vitro screening tools are essential in identifying promising compounds during drug development. Tests with currently used cell-based assays provide an indication of a compound's potential therapeutic benefits to the target tissue, but not to the whole body. Data obtained with animal models often cannot be extrapolated to humans. Multicompartment microfluidic-based devices, particularly those that are physical representations of physiologically based pharmacokinetic (PBPK) models, may contribute to improving the drug development process. These scaled-down devices, termed micro cell culture analogs (µCCAs) or body-on-a-chip devices, can simulate multitissue interactions under near-physiological fluid flow conditions and with realistic tissue-to-tissue size ratios. Because the device can be used with both animal and human cells, it can facilitate cross-species extrapolation. Used in conjunction with PBPK models, the devices permit an estimation of effective concentrations that can be used for studies with animal models or predict the human response. The devices also provide a means for relatively high-throughput screening of drug combinations and, when utilized with a patient's tissue sample, an opportunity for individualized medicine. Here we review efforts made toward the development of microfabricated cell culture systems and give examples that demonstrate their potential use in drug development, such as identifying synergistic drug interactions as well as simulating multiorgan metabolic interactions. In addition to their use in drug development, the devices also can be used to estimate the toxicity of chemicals as occupational hazards and environmental contaminants.

Fabrication and optimization of alginate hydrogel constructs for use in 3D neural cell culture.

Two-dimensional (2D) culture systems provide useful information about many biological processes. However, some applications including tissue engineering, drug transport studies, and analysis of cell growth and dynamics are better studied using three-dimensional (3D) culture systems. 3D culture systems can potentially offer higher degrees of organization and control of cell growth environments, more physiologically relevant diffusion characteristics, and permit the formation of more extensive 3D networks of cell-cell interactions. A 3D culture system has been developed using alginate as a cell scaffold, capable of maintaining the viability and function of a variety of neural cell types. Alginate was functionalized by the covalent attachment of a variety of whole proteins and peptide epitopes selected to provide sites for cell attachment. Alginate constructs were used to entrap a variety of neural cell types including astroglioma cells, astrocytes, microglia and neurons. Neural cells displayed process outgrowth over time in culture. Cell-seeded scaffolds were characterized in terms of their biochemical and biomechanical properties, effects on seeded neural cells, and suitability for use as 3D neural cell culture models.

Considerations for the design and construction of a synthetic platform cell for biotechnological applications.

The design and construction of an artificial bacterial cell could revolutionize biotechnological processes and technologies. A functional platform cell that can be easily customized for a pre-defined task would be useful for applications from producing therapeutics to decontaminating waste streams. The platform cell must be robust and highly efficient. A biotechnological platform cell is related to the concept of a minimal cell, but several factors beyond those necessary for a minimal cell must be considered for a synthetic organism designed for biotechnological applications. Namely, a platform cell must exhibit robust cell reproduction, decreased genetic drift, a physically robust cell envelope, efficient and simplified transcription and translation controls, and predictable metabolic interactions. Achieving a biotechnological platform cell will benefit from insights acquired from a minimal cell, but an approach of minimizing an existing organism's genome may be a more practical experimental approach. Escherichia coli possess many of the desired characteristics of a platform cell and could serve as a useful model organism for the design and construction of a synthetic platform organism. In this article we review briefly the current state of research in this field and outline specific characteristics that will be important for a biotechnologically relevant synthetic cell that has a minimized genome and efficient regulatory structure.

Incorporating genome-wide DNA sequence information into a dynamic whole-cell model of Escherichia coli: application to DNA replication.

The advent of thousands of annotated genomes, detailed metabolic reconstructions and databases within the flourishing field of systems biology necessitates the development of functionally complete computer models of whole cells and cellular systems. Such models would realistically describe fundamental properties of living systems such as growth, division and chromosome replication. This will inevitably bridge bioinformatic technologies with ongoing mathematical modelling efforts and would allow for in silico prediction of important dynamic physiological events. To demonstrate a potential for the anticipated merger of bioinformatic genome-wide data with a whole-cell computer model, the authors present here an updated version of a dynamic model of Escherichia coli, including a module that correctly describes the initiation and control of DNA replication by nucleoprotein DnaA-ATP molecules. Specifically, a rigorous mathematical approach used to explicitly include the genome-wide distribution of DnaA-binding sites on the replicating chromosome into a computer model of a bacterial cell is discussed. A new simple deterministic approximation of the complex stochastic process of DNA replication initiation is also provided. It is shown for the first time that reasonable assumptions about the mechanism of DNA replication initiation can be implemented in a deterministic whole-cell model to make predictions about the timing of chromosome replication. Furthermore, it is proposed that a large increase in the concentration of DnaA-binding boxes will result in a decreased steady-state growth rate in E. coli.

Three-dimensional hydrogel cultures for modeling changes in tissue impedance around microfabricated neural probes.

One limitation to the use of neuroprosthestic devices for chronic application, in the treatment of disease, is the reactive cell responses that occur surrounding the device after insertion. These cell and tissue responses result in increases in device impedance and failure of the device to interact with target populations of neurons. However, few tools are available to assess which components of the reactive response contribute most to changes in tissue impedance. An in vitro culture system has been developed that is capable of assessing individual components of the reactive response. The system utilizes alginate cell encapsulation to construct three-dimensional architectures that approach the cell densities found in rat cortex. The system was constructed around neuroNexus acute probes with on-board circuitry capable of monitoring the electrical properties of the surrounding tissue. This study demonstrates the utility of the system by demonstrating that differences in cell density within the three-dimensional alginate constructs result in differences in resistance and capacitance as measured by electrochemical impedance spectroscopy. We propose that this system can be used to model components of the reactive responses in brain tissue, and that the measurements recorded in vitro are comparable to measurements recorded in vivo.

A modular minimal cell model: purine and pyrimidine transport and metabolism.

A more complete understanding of the relationship of cell physiology to genomic structure is desirable. Because of the intrinsic complexity of biological organisms, only the simplest cells will allow complete definition of all components and their interactions. The theoretical and experimental construction of a minimal cell has been suggested as a tool to develop such an understanding. Our ultimate goal is to convert a "coarse-grain" lumped parameter computer model of Escherichia coli into a genetically and chemically detailed model of a "minimal cell." The base E. coli model has been converted into a generalized model of a heterotrophic bacterium. This coarse-grain minimal cell model is functionally complete, with growth rate, composition, division, and changes in cell morphology as natural outputs from dynamic simulations where only the initial composition of the cell and of the medium are specified. A coarse-grain model uses pseudochemical species (or modules) that are aggregates of distinct chemical species that share similar chemistry and metabolic dynamics. This model provides a framework in which these modules can be "delumped" into chemical and genetic descriptions while maintaining connectivity to all other functional elements. Here we demonstrate that a detailed description of nucleotide precursors transport and metabolism is successfully integrated into the whole-cell model. This nucleotide submodel requires fewer (12) genes than other theoretical predictions in minimal cells. The demonstration of modularity suggests the possibility of developing modules in parallel and recombining them into a fully functional chemically and genetically detailed model of a prokaryote cell.

A simple method for enhancing paclitaxel release from Taxus canadensis cell suspension cultures utilizing cell wall digesting enzymes.

Paclitaxel storage in Taxus suspension cell cultures was studied through the simple use of cell wall digesting enzymes. The application of cellulase (1%) and pectolyase (0.1%) to Taxus canadensis suspension cultures induced a significant increase in the paclitaxel present in the extracellular medium while maintaining membrane integrity, suggesting that paclitaxel is stored in the cell wall. The addition of cell wall digesting enzymes to a cell culture bioprocess may be an effective way of enhancing paclitaxel release to the extracellular medium and hence simplify product recovery.

Towards the development of a minimal cell model by generalization of a model of Escherichia coli: use of dimensionless rate parameters.

A model of a minimal cell would be a valuable tool in identifying the organizing principles that relate the static sequence information of the genome to the dynamic functioning of the living cell. Our approach for developing a minimal cell model is to first generalize an existing model of Escherichia coli by expressing reaction rates as ratios to a set of reference parameters. This generalized model is a prototype minimal cell model that will be developed by adding detail to explicitly include each chemical species. We tested the concept of a generalized model by testing the effect of scaling all enzyme-catalyzed reactions in the E. coli model. The scaling has little effect on cellular function for a wide range of kinetic ratios, where the kinetic ratio is defined as the rate of all enzyme-catalyzed reactions in a given model relative to those in the E. coli model.

Production of a sialylated N-linked glycoprotein in insect cells.

Under High Aspect Ratio Vessel (HARV) bioreactor culture conditions designed to simulate the microgravity of orbital space flight, insect tissue culture cells infected with a baculovirus expression vector produced a human glycoprotein with tri- and tetra-antennary complex N-linked oligosaccharides containing terminal sialic acid residues. The oligosaccharide structures were similar to those produced in human placental cells. Insect cells cultured in T-flasks only performed incomplete oligosaccharide processing. The mechanism of HARV-mediated changes in the eukaryotic N-linked glycosylation pathway was investigated and could be mimicked under T-flask growth conditions with the addition of N-acetylmannosamine to the culture medium. The significance of these investigations is discussed with respect to the production of recombinant therapeutic glycoproteins, insect physiology, and orbital space flight.

The effects of pH and surface composition on Pb adsorption to natural freshwater biofilms.

Two dominant variables that control the adsorption of toxic trace metals to suspended particulate materials and aquatic surface coatings are surface composition and solution pH. A model for the pH-dependent adsorption of Pbto heterogeneous particulate surface mixtures was derived from experimental evaluation of Pb adsorption to laboratory-derived surrogates. The surrogate materials were selected to represent natural reactive surface components. Pb adsorption to both the laboratory surrogates and natural biofilms was determined in chemically defined solutions under controlled laboratory conditions. Pb adsorption was measured over a pH range of 5-8, with an initial Pb concentration in solution of 2.0 microM. The surface components considered include amorphous Fe oxide, biogenic Mn oxide produced by a Mn(II) oxidizing bacterium (Leptothrix discophora SS-1), Al oxide, the common green alga Chlorella vulgaris, and Leptothrix discophora SS-1 cells. A linearization of Pb adsorption data for each adsorbent was used to quantify the relationship between Pb adsorption and pH. The parameters for individual adsorbents were incorporated into an additive model to predict the total Pb adsorption in multiple-adsorbent natural surface coatings that were collected from Cayuga Lake, NY. Pb adsorption experiments on the natural surface coatings at variable pH were utilized to verify the additive model predictions based on the pH dependent behavior of the experimental laboratory surrogates. Observed Pb adsorption is consistent with the model predictions (within 1-24%) over the range of solution pH values considered. The experimental results indicate that the combination of Fe and biogenic Mn oxides can contribute as much as 90% of Pb adsorbed on Cayuga Lake biofilms, with the dominant adsorbent switching from Mn to Fe oxide with increasing pH.

Comparison of growth and recombinant protein expression in two different insect cell lines in attached and suspension culture.

Culture conditions required for obtaining maximum recombinant protein concentrations from two cell lines, Spodoptera frugiperda (IPLBeta-Sf21-AE) and Trichoplusia ni (Tn 5Beta-1-4), were determined in this work. Conditions studied include mode of culture (suspended vs attached), agitation rates, inoculum sizes, cell concentration at the time of infection, and various serum-free media (SFM). Results were compared with the performance of attached cultures in TnM-FH with 10% fetal bovine serum. Growth rates in the different culture media tested were similar, but the cell numbers achieved (i.e., yield) improved 2 to 2.7-fold in SFM over cultures in TnM-FH. Agitation rates of 150-160 rpm were necessary for maximum growth of suspended Tn 5Beta-1-4 cells compared to 125-150 rpm for Sf-21 cells. An inoculum size of 5 x 10(5) cells/mL gave good growth rates and optimum biomass yields for both cell lines. Cultures of both cell lines were infected with viruses encoding for beta-galactosidase or human secreted alkaline phosphatase (seAP). Protein expression in TnM-FH in attached culture showed that Tn 5Beta-1-4 cells are 2-4.5 times more productive on a per cell basis than Sf-21 cells grown under similar conditions. Production of beta-galactosidase in Sf-21 cells increased 50% in suspension cultures with SFM compared to attached cultures in TnM-FH, but seAP expression was essentially unchanged by culture techniques. The Tn 5Beta-1-4 cells produced 2.6-4.4 and 2.7-3 times more beta-galactosidase and seAP, respectively, in SFM in suspension compared to Sf-21 cells. EX-CELL 401 and Sf900-II were formulated as optimized SFM for Sf cell lines. However, in Sf-21 cultures EX-CELL 400 performed better than the other two media, as it increased the beta-galactosidase yield up to 25%. Surprisingly, EX-CELL 401 was the best medium for the production of beta-galactosidase by Tn 5Beta-1-4 cells, resulting in 25% and 69% higher volumetric and specific yields, respectively, compared to EX-CELL 405 which was formulated for this specific cell line. These results show that even when culture media are designed for maximal growth of a specific cell line, other media may provide the best conditions for protein production.

Phenanthrene desorption from soil in the presence of bacterial extracellular polymer: observations and model predictions of dynamic behavior.

The extracellular polymer produced by a bacterium isolated from soil was employed in laboratory studies of desorption of a model polynuelear aromatic hydrocarbon (PAH), phenanthrene. The experimental results show that the selected extracellular polymer enhances the extent of release of soil-bound phenanthrene. A kinetic model was developed as an aid in interpreting the alterations in phenanthrene desorption resulting from polymer addition. The model employs a statistical gamma (gamma) distribution to describe spectrum of rate constants for transfer of phenanthrene from soil to water, and assumes instantaneous binding of phenanthrene to polymer and of polymer to the test soil. The relevant distribution coefficients and statistical parameters of the gamma distribution needed for the model were evaluated in independent experiments. Using these measured parameters, the model provides a satisfactory independent prediction of phenanthrene release from soil to aqueous phase at two test polymer concentrations, 50 mg TOC/L and 100 mg TOC/L. The success of the independent model predictions suggests a mechanism for the influence of extracellular polymer on phenanthrene desorption. The intrinsic, soil-specific, rate constants for solid to solution transfer of phenanthrene do not appear to be changed by bacterial polymer. Instead, polymer binding of phenanthrene in solution results in an increase in driving force for desorption by decreasing the solution concentration of the free, unbound, PAH molecule.

Cell science and protein crystal growth research for the International Space Station.

The recent National Research Council report, Future Biotechnology Research on the International Space Station, evaluates NASA's plans for research in cell science and protein crystal growth to be conducted on the International Space Station. This report concludes that the NASA biotechnology programs have the potential to significantly impact relevant scientific fields and to increase understanding and insight into fundamental biological issues. In order to realize the potential impacts, NASA must focus its research programs by selecting specific questions related to gravitational forces' role in cell behavior and by using the microgravity environment as a tool to determine the structure of macromolecules with important biological implications. Given the time and volume constraints associated with space-based experiments, instrumentation to be used on the space station must be designed to maximize the productivity of researchers, and NASA's recruitment of investigators and support for space station experiments should aim to encourage and facilitate cutting-edge research.

Influence of baculovirus-host cell interactions on complex N-linked glycosylation of a recombinant human protein.

The conditions required for mammalian-type complex N-linked glycosylation of human proteins produced in insect cells with the baculovirus expression vector system were investigated. Marked alterations to N-linked glycosylation of human placental secreted alkaline phosphatase (SEAP) were observed with different baculovirus species, insect cell lines, and cell culture media. When a recombinant Autographa californica nucleopolyhedrovirus (AcMNPV) was used to produce SEAP in Trichoplusia ni (Tn-4h) cells cultured in serum-free medium, structural analyses indicated <1% hybrid and no complex oligosaccharides attached to SEAP, a typical result with the baculovirus expression vector system. However, when fetal bovine serum was added to the culture medium, 48 +/- 4% of the oligosaccharides were hybrid or complex (but asialylated) glycans. When a recombinant T. ni nucleopolyhedrovirus (TnSNPV) was similarly used to express SEAP in Tn-4h cells cultured in serum-containing medium, only 24 +/- 3% of the glycans contained terminal N-acetylglucosamine and/or galactose residues. In contrast, SEAP produced in Sf9 cells grown in serum-containing medium with AcMNPV contained <1% hybrid oligosaccharides and no complex oligosaccharides. The results illustrate that baculovirus type, host cell type, and the growth medium all have a strong influence on the glycosylation pathway in insect cells, resulting in significant alterations in structures and relative abundance of N-linked glycoforms. Although the addition of sialic acid residues to the SEAP glycans was not detected, possible approaches to obtain sialylated glycans are discussed.

Combining cell culture analogue reactor designs and PBPK models to probe mechanisms of naphthalene toxicity.

An alternative method of evaluating the toxicology of a chemical is to use cultured mammalian cells in a novel cell culture analogue reactor (CCA) together with a corresponding physiologically based pharmacokinetic model (PBPK). The PBPK is a mathematical model that divides the body into compartments representing organs, integrating the kinetic, thermodynamic, and anatomical parameters of the animal. The bioreactor is a physical replica of the PBPK; where the PBPK specifies an organ or tissue compartment, the bioreactor contains compartments with a corresponding cell type. The device is a continuous, dynamic system composed of multiple cell types that interact through a common circulating cell culture medium. The bioreactor and the model are coupled to evaluate the plausibility of the molecular mechanism that is input into the model. This concept is tested with naphthalene as a model of PAH (polycyclic aromatic hydrocarbons) toxicants. Two physically different CCA reactors were tested with naphthalene, and different results were observed. In the prototype system using cells attached to glass dilution bottles, naphthalene dosing resulted in generation of a circulating metabolite from the "liver" compartment (based on H4IIE cells from a rat hepatoma) that caused cell death in the "lung" compartment (L2 cells from a rat lung), as well as depletion of glutathione in the L2 cells. An improved CCA using packed bed reactors of microcarrier cultured cells did not show differences between naphthalene-dosed and nondosed controls. To explain the different responses of the two CCA designs, PBPKs of the two reactors were tested with variations in physical and kinetic parameters, and toxic mechanism. When the toxic metabolite of naphthalene was naphthoquinone rather than naphthalene epoxide as initially assumed, the PBPK results were consistent with the results of the two CCA designs. This result indicates that the mechanism of naphthalene toxicity in the CCAs may be mediated through naphthoquinone formation. The CCA-PBPK concept is demonstrated to be applicable to the study of toxic mechanisms. In particular, use of this approach suggests that in vitro naphthalene toxicity is mediated through the naphthoquinone metabolite.

Process options in hepatitis B surface antigen extraction from transgenic potato.

The process conditions for recombinant hepatitis B surface antigen (HBsAg) extraction from transgenic potato were examined. The effects of temperature, the reducing agent beta-mercaptoethanol (BME), and proteinase inhibitors on the level of antigenic activity of recovered HBsAg were determined. Sedimentation profiles were performed to characterize HBsAg assembly into virus-like particles. Increasing the temperature of the sample for about 1 min increased the measured HBsAg antigenic activity. The optimum temperature was around 50 degrees C. A 3-fold enhancement of the antigenic activity was obtained in extract from transgenic potato expressing HBsAg, when monoclonal antibodies were used to assay for HBsAg. When antigenic activity was determined by polyclonal antibodies, no enhancement in the antigenic activity was obtained. Temperature may affect the conformation of the a epitope to which the monoclonal antibodies bind or alter the fluidity of surface lipid regions. BME increased the antigenic activity of HBsAg up to 4-fold when monoclonal antibodies directed against the a determinant were used, but there was no increase with polyclonal antibodies. This observation suggests that BME affects the structure or presentation of the a epitope. In the presence of BME and leupeptin, a proteinase inhibitor, higher antigenic activity was obtained. Leupeptin might protect the antigen, which might become more susceptible to proteolytic degradation after reduction, as a result of stimulation of sulfhydryl proteases. Although both temperature and BME increased the antigenic activity of HBsAg individually, when combined their interaction was antagonistic, resulting in reduced antigenic activity. Different proteinase inhibitors, including leupeptin, aprotinin, E-64, pefabloc, and pepstatin, had no significant effect on HBsAg from potato extract in a 2 h period in the absence of BME. The sedimentation profile of potato-produced HBsAg was determined in 5-30% sucrose gradients. Yeast-derived recombinant HBsAg was used as a positive control. The HBsAg from transgenic potato showed sedimentation and density properties that are very similar to the yeast-produced antigen, indicating assembly into virus-like particles. BME treatment did not change the sedimentation profile.

Characterization of a perfusion reactor utilizing mammalian cells on microcarrier beads.

Our overall objective is to develop a cell culture analogue bioreactor (CCA) that can be used together with a corresponding physiologically based pharmacokinetic model (PBPK) to evaluate molecular mechanisms of toxicity. The PBPK is a mathematical model that divides the body into compartments representing organs, integrating the kinetic, thermodynamic, and anatomical parameters of the animal. The CCA bioreactor is a physical replica of the PBPK; where the PBPK specifies organs, the CCA bioreactor contains compartments with a corresponding cell type that mimics some of the characteristic metabolism of that organ. The device is a continuous, dynamic system composed of multiple cell types that interact through a common circulating cell culture medium. The CCA bioreactor and the model can be coupled to evaluate the plausibility of the molecular mechanism that is input into the model. This paper focuses on the design, development, and characterization of a CCA bioreactor to be used in naphthalene dose response studies. A CCA bioreactor prototype developed previously is improved upon by culturing the cells on microcarrier beads. Microcarrier beads with cells attached can form packed beds with cell culture medium perfusing the beds. In this study, two packed beds of cells, one with L2 cells (rat lung) and one with H4IIE cells (rat hepatoma), are linked in a physiologically relevant arrangement by a common recirculating cell culture medium. Studies of this CCA bioreactor presented here include mixing profiles, effect of reactor environment on cell viability and intracellular glutathione, naphthalene distribution profile, and initial naphthalene dosing studies. Unlike the prototype system there is no detectable response to naphthalene addition; in a companion paper we show that this discrepancy can be explained by differences in liquid residence times in the organ compartments. The perfusion reactor design is shown to have significant operating improvements over prototype designs.

Computer model for glucose-limited growth of a single cell of Escherichia coli B/r-A. Reprinted from Biotechnology and Bioengineering, Vol. 26, Issue 3, Pp 203-216 (1984).

A computer model is described which is capable of predicting changes in cell composition, cell size, cell shape, and the timing of chromosome synthesis in response to changes in external glucose limitation. The model is constructed primarily from information on unrestricted growth in glucose minimal medium. The ability of the model to make reasonable quantitative predictions under glucose-limitation is a test of the plausibility of the basic biochemical mechanisms included in the model. Such a model should be of use in differentiating among competing hypotheses for biological mechanisms and in suggesting as yet unobserved phenomena. The last two points are illustrated with the testing of a mechanism for the control of the initiation of DNA synthesis and predictions on cell-width variations during the division cycle.

The effect of inoculum density and conditioned medium on the production of ajmalicine and catharanthine from immobilized Catharanthus roseus cells.

The effect of the cell-inoculum size and the addition of conditioned medium on ajmalicine and catharanthine production were studied using immobilized Catharanthus roseus cells. Higher specific-uptake rates of ammonium, nitrate, and sugars were observed in the low-inoculum-density cultures (50 g FW/L) compared to the high-inoculum-density cultures (100 g FW/L). Alkaloid production was not correlated with the exhaustion of a particular nutrient from the medium. The high-inoculum-density cultures produced higher ajmalicine concentrations throughout the experiment. Catharanthine production was similar between the two inoculum-density cultures. The addition of conditioned medium to MS-production medium dramatically improved the production of ajmalicine and catharanthine. The addition of conditioned medium enhanced ajmalicine production from immobilized Catharanthus roseus cultures on day 15 by at least two- to fourfold compared to media without the conditioning factors. Catharanthine production was increased by nearly fivefold in cultures with conditioned medium compared to those without conditioned medium. The enhancing effects of conditioned medium on alkaloid production were attributed to an unidentified factor produced and secreted by suspension cultures of C. roseus. The presence of conditioned medium also decreased the sucrose hydrolysis rate. The ajmalicine concentration in these immobilized cell cultures was found to be a function of the fresh-weight concentration, irrespective of the inoculum density or the culture medium. The medium choice and the inoculum density determined how rapidly fresh weight was accumulated and thus, how quickly ajmalicine was produced. Ajmalicine production correlated positively with fresh-weight concentration, but catharanthine production was not correlated with fresh-weight concentration.

Isolation of labeled 9-dihydrobaccatin III and related taxoids from cell cultures of taxuscell cultures of taxus canadensis elicited with m.

Cell suspension cultures of Taxus canadensis rapidly produced paclitaxel (1) and other taxoids in response to elicitation with methyl jasmonate. Three of these taxoids, of potential value in the synthesis of taxoid analogues, have been isolated from cell cultures of Taxus canadensis and identified as 13-acetyl-9-dihydrobaccatin III (2), baccatin VI (3), and 9-dihydrobaccatin III (4). Of these metabolites, 9-dihydrobaccatin III (4) has not been isolated from any Taxus species, whereas 13-acetyl-9-dihydrobaccatin III (2) and baccatin VI (3) have been isolated from a number of natural sources. 2D NMR techniques, mass spectrometry, and partial synthesis were used to rigorously elucidate the structure and stereochemistry of these natural products.