Mapping the molecular basis for growth related phenotypes in industrial producer CHO cell lines using differential proteomic analysis.

BACKGROUND: The ability to achieve high peak viable cell density earlier in CHO cell culture and maintain an extended cell viability throughout the production process is highly desirable to increase recombinant protein yields, reduce host cell impurities for downstream processing and reduce the cost of goods. In this study we implemented label-free LC-MS/MS proteomic profiling of IgG4 producing CHO cell lines throughout the duration of the cell culture to identify differentially expressed (DE) proteins and intracellular pathways associated with the high peak viable cell density (VCD) and extended culture VCD phenotypes. RESULTS: We identified key pathways in DNA replication, mitotic cell cycle and evasion of p53 mediated apoptosis in high peak VCD clonally derived cell lines (CDCLs). ER to Golgi vesicle mediated transport was found to be highly expressed in extended culture VCD CDCLs while networks involving endocytosis and oxidative stress response were significantly downregulated. CONCLUSION: This investigation highlights key pathways for targeted engineering to generate desirable CHO cell phenotypes for biotherapeutic production.

Levels of ADAM10 are reduced in Alzheimer's disease CSF.

BACKGROUND: The disintegrin metalloproteinase 10 (ADAM10) is the main α-secretase acting in the non-amyloidogenic processing of the amyloid precursor protein. This study assesses whether ADAM10 is present in cerebrospinal fluid (CSF), and whether it has potential as a biomarker for Alzheimer's disease (AD). METHODS: ADAM10 was characterized in human CSF samples by immunoprecipitation and western blotting using antibodies specific for different domains of the protein and by ultracentrifugation in sucrose density gradients. Samples from AD patients (n = 20) and age-matched non-AD controls (n = 20) were characterized for classical CSF biomarkers, Aß42, T-tau, or P-tau by ELISA, and assayed for soluble ADAM10 levels by western blotting. RESULTS: We found that ADAM10 is present in human CSF as several distinct species: an immature form retaining the prodomain (proADAM10; ~ 80 kDa), a mature unprocessed full-length form (ADAM10f; ~ 55 kDa), and a truncated large soluble form released from the membrane (sADAM10; ~ 50 kDa). Fractionation by ultracentrifugation on sucrose density gradients showed that the ADAM10f and sADAM10 species form large complexes. Immunoblotting revealed a significant decrease in ADAM10f and sADAM10 in AD CSF compared to control CSF, while proADAM10 levels remained unaltered. CONCLUSIONS: Several forms of ADAM10 are present in CSF, mainly assembled as high-molecular weight complexes. The determination of the levels of mature forms of CSF-ADAM10 may be useful as a biomarker for AD.

Proteomics in biomanufacturing control: Protein dynamics of CHO-K1 cells and conditioned media during apoptosis and necrosis.

Cell viability has a critical impact on product quantity and quality during the biomanufacturing of therapeutic proteins. An advanced understanding of changes in the cellular and conditioned media proteomes upon cell stress and death is therefore needed for improved bioprocess control. Here, a high pH/low pH reversed phase data independent 2D-LC-MSE discovery proteomics platform was applied to study the cellular and conditioned media proteomes of CHO-K1 apoptosis and necrosis models where cell death was induced by staurosporine exposure or aeration shear in a benchtop bioreactor, respectively. Functional classification of gene ontology terms related to molecular functions, biological processes, and cellular components revealed both cell death independent and specific features. In addition, label free quantitation using the Hi3 approach resulted in a comprehensive shortlist of 23 potential cell viability marker proteins with highest abundance and a significant increase in the conditioned media upon induction of cell death, including proteins related to cellular stress response, signal mediation, cytoskeletal organization, cell differentiation, cell interaction as well as metabolic and proteolytic enzymes which are interesting candidates for translating into targeted analysis platforms for monitoring bioprocessing response and increasing process control.

Optimizing cell-free protein expression in CHO: Assessing small molecule mass transfer effects in various reactor configurations.

Cell-free protein synthesis (CFPS) is an ideal platform for rapid and convenient protein production. However, bioreactor design remains a critical consideration in optimizing protein expression. Using turbo green fluorescent protein (tGFP) as a model, we tracked small molecule components in a Chinese Hamster Ovary (CHO) CFPS system to optimize protein production. Here, three bioreactors in continuous-exchange cell-free (CECF) format were characterized. A GFP optical sensor was built to monitor the product in real-time. Mass transfer of important substrate and by-product components such as nucleoside triphosphates (NTPs), creatine, and inorganic phosphate (Pi) across a 10-kDa MWCO cellulose membrane was calculated. The highest efficiency measured by tGFP yields were found in a microdialysis device configuration; while a negative effect on yield was observed due to limited mass transfer of NTPs in a dialysis cup configuration. In 24-well plate high-throughput CECF format, addition of up to 40 mM creatine phosphate in the system increased yields by up to ∼60% relative to controls. Direct ATP addition, as opposed to creatine phosphate addition, negatively affected the expression. Pi addition of up to 30 mM to the expression significantly reduced yields by over ∼40% relative to controls. Overall, data presented in this report serves as a valuable reference to optimize the CHO CFPS system for next-generation bioprocessing. Biotechnol. Bioeng. 2017;114: 1478-1486. © 2017 Wiley Periodicals, Inc.

High throughput chromatography strategies for potential use in the formal process characterization of a monoclonal antibody.

High throughput experimental strategies are central to the rapid optimization of biologics purification processes. In this work, we extend common high throughput technologies towards the characterization of a multi-column chromatography process for a monoclonal antibody (mAb). Scale-down strategies were first evaluated by comparing breakthrough, retention, and performance (yields and clearance of aggregates and host cell protein) across miniature and lab scale columns. The process operating space was then evaluated using several integrated formats, with batch experimentation to define process testing ranges, miniature columns to evaluate the operating space, and comparison to traditional scale columns to establish scale-up correlations and verify the determined operating space. When compared to an independent characterization study at traditional lab column scale, the high throughput approach identified the same control parameters and similar process sensitivity. Importantly, the high throughput approach significantly decreased time and material needs while improving prediction robustness. Miniature columns and manufacturing scale centerpoint data comparisons support the validity of this approach, making the high throughput strategy an attractive and appropriate scale-down tool for the formal characterization of biotherapeutic processes in the future if regulatory acceptance of the miniature column data can be achieved. Biotechnol. Bioeng. 2016;113: 1273-1283. © 2015 Wiley Periodicals, Inc.

Host cell protein impurities in chromatographic polishing steps for monoclonal antibody purification.

Downstream purification of monoclonal antibodies (mAbs) is normally performed using a platform process that is empirically tuned to optimize impurity removal for each new product. A more fundamental understanding of impurities and the product itself would provide insights into the rational design of efficient downstream processes. This work examines the chromatographic properties of Chinese hamster ovary host cell protein (HCP) impurities in non-affinity chromatographic resins commonly used in polishing steps for monoclonal antibody purification: ion-exchange, hydrophobic interaction, and multimodal. Using proteomic analysis, the specific HCP impurities that elute close to mAb products are identified for these resins at typical downstream processing conditions. Additionally, the interactions of HCP impurities with mAb products are profiled to determine the total extent of product association and the specific HCP species that form associative complexes under conditions encountered in polishing columns. Product association and co-elution were both identified as viable mechanisms of HCP retention for the non-affinity resins tested here. A relatively large sub-population of HCP impurities was found to co-elute or associate with mAbs in each polishing column, but only a small population of HCPs-including lipoprotein lipase, chrondroitin sulfate proteoglycan 4, nidogen-1, and SPARC-were identified as difficult to remove across an entire downstream mAb process. Biotechnol. Bioeng. 2016;113: 1260-1272. © 2015 Wiley Periodicals, Inc.

The distribution pattern of α2,3- and α2,6-linked sialic acids affects host cell preference in Toxoplasma gondii.

Tachyzoites of Toxoplasma gondii, an obligate intracellular parasite, actively invade almost all types of nucleated cells. However, T. gondii tachyzoites preferentially infect particular types of animal tissue cells. The mechanism underlying the host cell preference of T. gondii is not yet known. In this study, we found that enzymatic removal of α2,3- but not α2,6-linked sialic acids on the surface of Vero cells decreased T. gondii tachyzoite adhesion or invasion to the treated cells. Although Chinese hamster ovary (CHO) cells express only α2,3-linked sialic acid, a genetically modified CHO cell line constructed by transfection with the α2,6-sialiltransferase gene contains subpopulations with a variety of expression patterns of α2,3- and α2,6-linked sialic acids. When T. gondii tachyzoites were added to the modified CHO cells, the tachyzoites preferentially attached to cells belonging to a subpopulation of cells that highly expressed α2,3-linked sialic acids. Additionally, multiple regression analysis performed to analyse the relationship between the amount of α2,3-linked/α2,6-linked sialic acids and parasite-expressed fluorescence intensity suggested that more tachyzoites adhered to individual α2,3-linked sialic acid rich-cells than to α2,3-linked sialic acid-poor/null cells. The results of confocal laser microscopy confirmed this finding. These results indicate that the host cell preference of T. gondii was, at least partially, affected by the distribution pattern of α2,3-, but almost never α2,6-linked sialic acids on host cells.

MER5101, a novel Aß1-15:DT conjugate vaccine, generates a robust anti-Aß antibody response and attenuates Aß pathology and cognitive deficits in APPswe/PS1ΔE9 transgenic mice.

Active amyloid-ß (Aß) immunotherapy is under investigation to prevent or treat early Alzheimer's disease (AD). In 2002, a Phase II clinical trial (AN1792) was halted due to meningoencephalitis in ∼6% of the AD patients, possibly caused by a T-cell-mediated immunological response. Thus, generating a vaccine that safely generates high anti-Aß antibody levels in the elderly is required. In this study, MER5101, a novel conjugate of Aß1-15 peptide (a B-cell epitope fragment) conjugated to an immunogenic carrier protein, diphtheria toxoid (DT), and formulated in a nanoparticular emulsion-based adjuvant, was administered to 10-month-old APPswe/PS1ΔE9 transgenic (Tg) and wild-type (Wt) mice. High anti-Aß antibody levels were observed in both vaccinated APPswe/PS1ΔE9 Tg and Wt mice. Antibody isotypes were mainly IgG1 and IgG2b, suggesting a Th2-biased response. Restimulation of splenocytes with the Aß1-15:DT conjugate resulted in a strong proliferative response, whereas proliferation was absent after restimulation with Aß1-15 or Aß1-40/42 peptides, indicating a cellular immune response against DT while avoiding an Aß-specific T-cell response. Moreover, significant reductions in cerebral Aß plaque burden, accompanied by attenuated microglial activation and increased synaptic density, were observed in MER5101-vaccinated APPswe/PS1ΔE9 Tg mice compared with Tg adjuvant controls. Last, MER5101-immunized APPswe/PS1ΔE9 Tg mice showed improvement of cognitive deficits in both contextual fear conditioning and the Morris water maze. Our novel, highly immunogenic Aß conjugate vaccine, MER5101, shows promise for improving Aß vaccine safety and efficacy and therefore, may be useful for preventing and/or treating early AD.

Metabolite profiling of recombinant CHO cells: designing tailored feeding regimes that enhance recombinant antibody production.

Chinese hamster ovary (CHO) cells are the primary platform for commercial expression of recombinant therapeutic proteins. Obtaining maximum production from the expression platform requires optimal cell culture medium (and associated nutrient feeds). We have used metabolite profiling to define the balance of intracellular and extracellular metabolites during the production process of a CHO cell line expressing a recombinant IgG4 antibody. Using this metabolite profiling approach, it was possible to identify nutrient limitations, which acted as bottlenecks for antibody production, and subsequently develop a simple feeding regime to relieve these metabolic bottlenecks. This metabolite profiling-based strategy was used to design a targeted, low cost nutrient feed that increased cell biomass by 35% and doubled the antibody titer. This approach, with the potential for utilization in non-specialized laboratories, can be applied universally to the optimization of production of commercially important biopharmaceuticals.

Rapid characterization of protein productivity and production stability of CHO cells by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry.

The biopharmaceutical industry has been in pursuit of strategies which can isolate stable and high-producing cell lines. The whole cell mass spectrometry method by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) is a rapid and simple method for cell characterization based on the differences in the fingerprints of the mass spectra. This work describes how the method was evaluated for the application of screening for stable and high-producing clones from a panel of recombinant Chinese hamster ovary (CHO) cell lines. Detectable m/z values and their relative intensities were collected and processed by partial least squares (PLS). To reduce the errors introduced by the preparation method and spectra noise, high intensity preliminary data was selected and the number of variables introduced was validated by leave-one-out cross-validation. The differences in recombinant protein productivity and titer were revealed by PLS regression with promising results. Partial least-squares discriminant analysis (PLS-DA) was applied to differentiate stable and unstable cell lines as traditional stability testing would require several months involving numerous continuous passages. Results confirmed that the whole cell MALDI-TOF method can be a powerful method for routine monitoring of bioprocesses and study can be further developed by extending the number of the cell lines tested to establish a recombinant cell line database.

Investigating the influence of physiologically relevant hydrostatic pressure on CHO cell batch culture.

Chinese hamster ovary (CHO) cells have been the most commonly used mammalian host for large-scale commercial production of therapeutic proteins, such as monoclonal antibodies. Enhancement of productivity of these CHO cells is one of the top priorities in the biopharmaceutical industry to reduce manufacturing cost. Although there are many different methods (e.g. temperature, pH, feed) to improve protein production in CHO cells, the role of physiologically relevant hydrostatic pressure in CHO cell culture has not been reported yet. In this study, four different hydrostatic pressures (0, 30, 60, and 90 mmHg) were applied to batch CHO cells, and their cell growth/metabolism and IgG1 production were examined. Our results indicate that hydrostatic pressure can increase the maximum cell concentration by up to 50%. Moreover, overall IgG1 concentration on Day 5 showed that 30 mmHg pressure can increase IgG1 production by 26%. The percentage of non-disulphide-linked antibody aggregates had no significant change under pressure. Besides, no significant difference was observed between 30 mmHg and no pressure conditions in terms of cell clumping formation. All these findings are important for the optimization of fed-batch or perfusion culture for directing cell growth and improving antibody production.

Rapid characterization of high/low producer CHO cells using matrix-assisted laser desorption/ionization time-of-flight.

An intact-cell mass spectrometry (ICM) method using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) was evaluated for the screening of stable recombinant Chinese hamster ovary (CHO) cell lines, an important mammalian cell line in bioprocessing. With rapid and simple cell pretreatments, viabilities of cells could be rapidly distinguished on the different fingerprints of mass spectra. Detectable m/z values on cell surfaces and their relative intensities were processed by two biostatistical methods, principle components analysis (PCA) and partial least squares (PLS), with promising results. Discrimination among cell lines with different expressed recombinant proteins or different productivities could be achieved. The ICM method has the advantage of providing multiple parameters simultaneously and possesses the potential to become a powerful method for routine monitoring of bioprocesses.

Application of Raman spectroscopy in monoclonal antibody producing continuous systems for downstream process intensification.

Monoclonal antibodies (mAbs) are biopharmaceuticals produced by mammalian cell lines in bioreactors at a variety of scales. Cell engineering, media optimization, process monitoring, and control strategies for in vitro production have become crucial subjects to meet increasing demand for these high value pharmaceuticals. Raman Spectroscopy has gained great attention in the pharmaceutical industry for process monitoring and control to maintain quality assurance. For the first time, this article demonstrated the possibility of subclass independent quantitative mAb prediction by Raman spectroscopy in real time. The developed model estimated the concentrations of different mAb isotypes with average prediction errors of 0.2 (g/L) over the course of cell culture. In situ Raman spectroscopy combined with chemometric methods showed to be a useful predictive tool for monitoring of real time mAb concentrations in a permeate stream without sample removal. Raman spectroscopy can, therefore, be considered as a reliable process analytical technology tool for process monitor, control, and intensification of downstream continuous manufacturing. The presented results provide useful information for pharmaceutical industries to choose the most appropriate spectroscopic technology for their continuous processes.

Improving Chinese hamster ovary host cell protein ELISA using Ella®: an automated microfluidic platform.

Chinese hamster ovary (CHO) cells are a mammalian cell line used in the production of therapeutic proteins. Host cell proteins (HCPs) are process-related impurities that are derived from the host cell expression system. During biopharmaceutical drug development, removal of HCPs is required. Enzyme-linked immunosorbent assay (ELISA) is a common technique to quantitate HCPs, but is a labor-intensive process that takes up to 7 h. Ella® is an automated instrument that utilizes microfluidics and glass nanoreactors to quantitate HCPs in 75 min using similar ELISA reagents. The antibodies and antigens are captured on three distinct glass nanoreactors, resulting in sensitive reproducible data. Our results indicate that Ella quantitates CHO HCPs with precision, accuracy, sensitivity and trends comparable with our traditional CHO HCP ELISA.

Analysis of chemometric models applied to Raman spectroscopy for monitoring key metabolites of cell culture.

The Food and Drug Administration (FDA) initiative of Process Analytical Technology (PAT) encourages the monitoring of biopharmaceutical manufacturing processes by innovative solutions. Raman spectroscopy and the chemometric modeling tool partial least squares (PLS) have been applied to this aim for monitoring cell culture process variables. This study compares the chemometric modeling methods of Support Vector Machine radial (SVMr), Random Forests (RF), and Cubist to the commonly used linear PLS model for predicting cell culture components-glucose, lactate, and ammonia. This research is performed to assess whether the use of PLS as standard practice is justified for chemometric modeling of Raman spectroscopy and cell culture data. Model development data from five small-scale bioreactors (2 × 1 L and 3 × 5 L) using two Chinese hamster ovary (CHO) cell lines were used to predict against a manufacturing scale bioreactor (2,000 L). Analysis demonstrated that Cubist predictive models were better for average performance over PLS, SVMr, and RF for glucose, lactate, and ammonia. The root mean square error of prediction (RMSEP) of Cubist modeling was acceptable for the process concentration ranges of glucose (1.437 mM), lactate (2.0 mM), and ammonia (0.819 mM). Interpretation of variable importance (VI) results theorizes the potential advantages of Cubist modeling in avoiding interference of Raman spectral peaks. Predictors/Raman wavenumbers (cm-1 ) of interest for individual variables are X1139-X1141 for glucose, X846-X849 for lactate, and X2941-X2943 for ammonia. These results demonstrate that other beneficial chemometric models are available for use in monitoring cell culture with Raman spectroscopy.

A cell-based assay for the detection of neutralizing antibodies against alemtuzumab.

Aim: The humanized anti-CD52 monoclonal antibody alemtuzumab depletes lymphocytes and is currently used to treat relapsing multiple sclerosis. During treatment, anti-alemtuzumab antibodies may develop and reduce effective lymphocyte depletion in future treatment cycles. Results: Alemtuzumab-Alexa Fluor 488 conjugate binding to the CHO-CD52 cell surface was inhibited by anti-alemtuzumab antibodies. Conclusion: In this proof-of-concept study, a CHO-CD52 cell line has been developed and used to detect the presence of anti-alemtuzumab neutralizing antibodies. This platform provides the basis of an assay for routine screening of serum for neutralizing antibodies from patients treated with alemtuzumab.

Expression of TRPC3 in Chinese hamster ovary cells results in calcium-activated cation currents not related to store depletion.

TRPC3 (or Htrp3) is a human member of the trp family of Ca2+-permeable cation channels. Since expression of TRPC3 cDNA results in markedly enhanced Ca2+ influx in response to stimulation of membrane receptors linked to phospholipase C (Zhu, X., J. Meisheng, M. Peyton, G. Bouley, R. Hurst, E. Stefani, and L. Birnbaumer. 1996. Cell. 85:661-671), we tested whether TRPC3 might represent a Ca2+ entry pathway activated as a consequence of depletion of intracellular calcium stores. CHO cells expressing TRPC3 after intranuclear injection of cDNA coding for TRPC3 were identified by fluorescence from green fluorescent protein. Expression of TRPC3 produced cation currents with little selectivity for Ca2+ over Na+. These currents were constitutively active, not enhanced by depletion of calcium stores with inositol-1,4,5-trisphosphate or thapsigargin, and attenuated by strong intracellular Ca2+ buffering. Ionomycin led to profound increases of currents, but this effect was strictly dependent on the presence of extracellular Ca2+. Likewise, infusion of Ca2+ into cell through the patch pipette increased TRPC3 currents. Therefore, TRPC3 is stimulated by a Ca2+-dependent mechanism. Studies on TRPC3 in inside-out patches showed cation-selective channels with 60-pS conductance and short (<2 ms) mean open times. Application of ionomycin to cells increased channel activity in cell-attached patches. Increasing the Ca2+ concentration on the cytosolic side of inside-out patches (from 0 to 1 and 30 microM), however, failed to stimulate channel activity, even in the presence of calmodulin (0.2 microM). We conclude that TRPC3 codes for a Ca2+-permeable channel that supports Ca2+-induced Ca2+-entry but should not be considered store operated.

Structure of synaptogyrin (p29) defines novel synaptic vesicle protein.

Synaptogyrin (p29) is a synaptic vesicle protein that is uniformly distributed in the nervous system (Baumert et al., 1990). We have cloned and sequenced the cDNA encoding synaptogyrin, and the sequence predicts a protein with a molecular mass of 25,900 D with four membrane-spanning domains. The topology of the protein was confirmed by limited proteolysis using domain-specific antibodies. Database searches revealed several cDNA sequences coding polypeptides with sequence identities ranging from 32 to 46%, suggesting that synaptogyrin is a member of a multigene family. When the synaptogyrin cDNA is expressed in COS cells, the generated protein is indistinguishable from native synaptogyrin. To study intracellular sorting, synaptogyrin was expressed in CHO cells that revealed a punctate staining that was very similar to that of synaptophysin and endogenously expressed cellubrevin. Significant overlap with transferrin staining was also observed, suggesting that synaptogyrin is targeted to a recycling compartment involved in membrane traffic to and from the plasma membrane.

Identification of LIM3 as the principal determinant of paxillin focal adhesion localization and characterization of a novel motif on paxillin directing vinculin and focal adhesion kinase binding.

Paxillin is a 68-kD focal adhesion phosphoprotein that interacts with several proteins including members of the src family of tyrosine kinases, the transforming protein v-crk, and the cytoskeletal proteins vinculin and the tyrosine kinase, focal adhesion kinase (FAK). This suggests a function for paxillin as a molecular adaptor, responsible for the recruitment of structural and signaling molecules to focal adhesions. The current study defines the vinculin- and FAK-interaction domains on paxillin and identifies the principal paxillin focal adhesion targeting motif. Using truncation and deletion mutagenesis, we have localized the vinculin-binding site on paxillin to a contiguous stretch of 21 amino acids spanning residues 143-164. In contrast, maximal binding of FAK to paxillin requires, in addition to the region of paxillin spanning amino acids 143-164, a carboxyl-terminal domain encompassing residues 265-313. These data demonstrate the presence of a single binding site for vinculin, and at least two binding sites for FAK that are separated by an intervening stretch of 100 amino acids. Vinculin- and FAK-binding activities within amino acids 143-164 were separable since mutation of amino acid 151 from a negatively charged glutamic acid to the uncharged polar residue glutamine (E151Q) reduced binding of vinculin to paxillin by >90%, with no reduction in the binding capacity for FAK. The requirement for focal adhesion targeting of the vinculin- and FAK-binding regions within paxillin was determined by transfection into CHO.K1 fibroblasts. Significantly and surprisingly, paxillin constructs containing both deletion and point mutations that abrogate binding of FAK and/or vinculin were found to target effectively to focal adhesions. Additionally, expression of the amino-terminal 313 amino acids of paxillin containing intact vinculin- and FAK-binding domains failed to target to focal adhesions. This indicated other regions of paxillin were functioning as focal adhesion localization motifs. The carboxyl-terminal half of paxillin (amino acids 313-559) contains four contiguous double zinc finger LIM domains. Transfection analyses of sequential carboxyl-terminal truncations of the four individual LIM motifs and site-directed mutagenesis of LIM domains 1, 2, and 3, as well as deletion mutagenesis, revealed that the principal mechanism of targeting paxillin to focal adhesions is through LIM3. These data demonstrate that paxillin localizes to focal adhesions independent of interactions with vinculin and/or FAK, and represents the first definitive demonstration of LIM domains functioning as a primary determinant of protein subcellular localization to focal adhesions.

Differential localization of VE- and N-cadherins in human endothelial cells: VE-cadherin competes with N-cadherin for junctional localization.

The two major cadherins of endothelial cells are neural (N)-cadherin and vascular endothelial (VE)- cadherin. Despite similar level of protein expression only VE-cadherin is located at cell-cell contacts, whereas N-cadherin is distributed over the whole cell membrane. Cotransfection of VE-cadherin and N-cadherin in CHO cells resulted in the same distribution as that observed in endothelial cells indicating that the behavior of the two cadherins was not cell specific but related to their structural characteristics. Similar amounts of alpha- and beta-catenins and plakoglobin were associated to VE- and N-cadherins, whereas p120 was higher in the VE-cadherin complex. The presence of VE-cadherin did not affect N-cadherin homotypic adhesive properties or its capacity to localize at junctions when cotransfectants were cocultured with cells transfected with N-cadherin only. To define the molecular domain responsible for the VE-cadherin-dominant activity we prepared a chimeric construct formed by VE-cadherin extracellular region linked to N-cadherin intracellular domain. The chimera lost the capacity to exclude N-cadherin from junctions indicating that the extracellular domain of VE-cadherin alone is not sufficient for the preferential localization of the molecule at the junctions. A truncated mutant of VE-cadherin retaining the full extracellular domain and a short cytoplasmic tail (Arg621-Pro702) lacking the catenin-binding region was able to exclude N-cadherin from junctions. This indicates that the Arg621-Pro702 sequence in the VE-cadherin cytoplasmic tail is required for N-cadherin exclusion from junctions. Competition between cadherins for their clustering at intercellular junctions in the same cell has never been described before. We speculate that, in the endothelium, VE- and N-cadherin play different roles; whereas VE-cadherin mostly promotes the homotypic interaction between endothelial cells, N-cadherin may be responsible for the anchorage of the endothelium to other surrounding cell types expressing N-cadherin such as vascular smooth muscle cells or pericytes.