Mesoscale Polarization by Geometric Frustration in Columnar Supramolecular Crystals.

Columnar supramolecular phases with polarization along the columnar axis have potential for the development of ultrahigh-density memories as every single column might function as a memory element. By investigating structure and disorder for four columnar benzene-1,3,5-trisamides by total X-ray scattering and DFT calculations, we demonstrate that the column orientation, and thus the columnar dipole moment, is receptive to geometric frustration if the columns aggregate in a hexagonal rod packing. The frustration suppresses conventional antiferroelectric order and heightens the sensitivity towards collective intercolumnar packing effects. The latter finding allows for the building up of mesoscale domains with spontaneous polarization. Our results suggest how the complex interplay between steric and electrostatic interactions is influenced by a straightforward chemical design of the molecular synthons to create spontaneous polarization and to adjust mesoscale domain size.

Influence of proton coupling on symmetry-based homonuclear (19)F dipolar recoupling experiments.

We study the efficiency of two symmetry based homonuclear (19)F double-quantum recoupling sequences for moderate (R142(6)) and ultra-fast (R144(5)) MAS under the influence of strong (1)H-(1)H and (1)H-(19)F dipolar interactions and (1)H continuous wave decoupling. Simulations based on various spin systems derived from the organic solid 1,3,5-tris(2-fluoro-2-methylpropionylamino)benzene (F-BTA), used as a model system, reveal that the strong-decoupling limit is not accessible even for moderate spinning speeds. Additionally, for the no-decoupling limit improved DQ efficiencies are predicted for both moderate and ultra-fast MAS. Strong perturbations of build-up curves can be avoided by additional stabilisation through supercycling. Additional (1)H cw decoupling during (19)F recoupling rapidly reduces the maximum DQ efficiency when deviating from the no-decoupling limit. These effects were confirmed by experimental data on F-BTA. For moderate spinning the influence of (1)H-(1)H and (1)H-(19)F couplings is markedly stronger compared to ultra-fast MAS. For the latter case those influences reduce to a constant scaling if only short excitation times up to the first minimum are taken into account. Based on this analysis the experimental build-up curves of 1,3,5-tris(2-fluoro-2-methylpropionylamino)benzene can be refined with homonuclear (19)F spin systems which allow to probe even subtle structural differences for the fluorine atoms of F-BTA.

Cell-based shear stress sensor for bioprocessing.

Shear stress during bioreactor cultivation has significant impact on cell health, growth, and fate. Mammalian cells, such as T cells and stem cells, in next-generation cell therapies are especially more sensitive to shear stress present in their culture environment than bacteria. Therefore, a base knowledge about the shear stress imposed by the bioprocesses is needed to optimize the process parameters and enhance cell growth and yield. However, typical computational flow dynamics modeling or PCR-based assays have several limitations. Implementing and interpreting computational modeling often requires technical specialties and also relies on many simplifications in modeling. PCR-based assays evaluating changes in gene expression involve cumbersome sample preparation with the use of advanced lab equipment and technicians, hampering rapid and straightforward assessment of shear stress. Here, we developed a simple, cell-based shear stress sensor for measuring shear stress levels in different bioreactor types and operating conditions. We engineered a CHO-DG44 cell line to make its stress sensitive promoter EGR-1 control GFP expression. Subsequently, the stressed CHO cells were transferred into a 96 well plate, and their GFP levels (population mean fluorescence) were monitored using a cell analysis instrument (Incucyte®, Sartorius Stedim Biotech) over 24 hours. After conducting sensor characterization, which included chemical induced stress and fluid shear stress, and stability investigation, we tested the shear stress sensor in the Ambr® 250 bioreactor vessels (Sartorius Stedim Biotech) with different impeller and vessel designs. The results showed that the CHO cell-based shear stress sensors expressed higher GFP levels in response to higher shear stress magnitude or exposure time. These sensors are useful tools to assess shear stress imposed by bioreactor conditions and can facilitate the design of various bioreactor vessels with a low shear stress profile.

Feasibility and performance of cross-clone Raman calibration models in CHO cultivation.

Raman spectroscopy is widely used in monitoring and controlling cell cultivations for biopharmaceutical drug manufacturing. However, its implementation for culture monitoring in the cell line development stage has received little attention. Therefore, the impact of clonal differences, such as productivity and growth, on the prediction accuracy and transferability of Raman calibration models is not yet well described. Raman OPLS models were developed for predicting titer, glucose and lactate using eleven CHO clones from a single cell line. These clones exhibited diverse productivity and growth rates. The calibration models were evaluated for clone-related biases using clone-wise linear regression analysis on cross validated predictions. The results revealed that clonal differences did not affect the prediction of glucose and lactate, but titer models showed a significant clone-related bias, which remained even after applying variable selection methods. The bias was associated with clonal productivity and lead to increased prediction errors when titer models were transferred to cultivations with productivity levels outside the range of their training data. The findings demonstrate the feasibility of Raman-based monitoring of glucose and lactate in cell line development with high accuracy. However, accurate titer prediction requires careful consideration of clonal characteristics during model development.

Optimized signal peptides for the development of high expressing CHO cell lines.

Recombinant biotherapeutic proteins such as monoclonal antibodies are mostly produced in Chinese hamster ovary (CHO) cells and pharmaceutical companies are interested in an appropriate platform technology for the development of large-scale production processes. A major aim of our study was therefore to improve the secretion efficiency of a recombinant biotherapeutic antibody by optimizing signal peptides. Reporter molecules such as gaussia and vargula luciferase or secreted alkaline phosphatase are frequently used to this end. In striking contrast, we used a biotherapeutic antibody that was fused to 16 different signal peptides during our study. In this way, the secretion efficiency of the recombinant antibody has been analyzed by transient expression experiments in CHO cell lines. Compared to the control signal peptide, it was not possible to achieve higher efficiencies with signal peptides derived from a variety of species or even natural immunoglobulin G signal peptides. The best results were obtained with natural signal peptides derived from human albumin and human azurocidin. These results were confirmed by fed-batch experiments with stably transfected cell pools, in which cell-specific productivities up to 90 pg cell(-1) day(-1) and product concentrations up to 4 g L(-1) could be determined using the albumin signal peptide. Finally, the applicability of the identified signal peptides for both different antibodies and non-antibody products was demonstrated by transient expression experiments. In conclusion, it was found that signal peptides derived from human albumin and human azurocidin are most appropriate to generate cell lines with clearly improved production rates suitable for commercial purposes in a product-independent manner.

Development of a novel ER stress based selection system for the isolation of highly productive clones.

Most biotherapeutic drugs are recombinant monoclonal antibodies which are mostly produced in monoclonal cell lines derived from Chinese hamster ovary (CHO) cells. Various clones expressing a monoclonal recombinant antibody were analyzed and a correlation of the antibody concentration and the relative mRNA level of calreticulin (CALR), glucose-regulated protein 78 and 94 kDa (GRP78, GRP94) and spliced X-box binding protein 1 (XPB1) was observed. By means of these results we were motivated to establish a novel selection system based on endoplasmic reticulum (ER) stress, which allows the rapid identification and isolation of high-expressing clones out of a pool mainly consisting of low- and medium-producing cells. Several ER stress responsive elements were tested with the aid of a recombinase mediated cassette exchange (RMCE) procedure. Very surprisingly, only GRP78 reporter constructs were strongly stimulated upon antibody expression. Furthermore we found that GRP78 reporter constructs are very suitable to reflect the level of antibody expression (IgG) in recombinant CHO cells. Based on these results, it is concluded, that the novel ER stress based selection system developed during this study is suitable to identify and isolate clones with a high level of antibody expression.

Hierarchical time-series analysis of dynamic bioprocess systems.

BACKGROUND: Monoclonal antibodies (mAbs) are leading types of 'blockbuster' biotherapeutics worldwide; they have been successfully used to treat various cancers and chronic inflammatory and autoimmune diseases. Biotherapeutics process development and manufacturing are complicated due to lack of understanding the factors that impact cell productivity and product quality attributes. Understanding complex interactions between cells, media, and process parameters on the molecular level is essential to bring biomanufacturing to the next level. This can be achieved by analyzing cell culture metabolic levels connected to vital process parameters like viable cell density (VCD). However, VCD and metabolic profiles are dynamic parameters and inherently correlated with time, leading to a significant correlation without actual causality. Many time-series methods deal with such issues. However, with metabolic profiling, the number of measured variables vastly exceeds the number of experiments, making most of existing methods ill-suited and hard to interpret. METHODS AND MAJOR RESULTS: Here we propose an alternative workflow using hierarchical dimension reduction to visualize and interpret the relation between evolution of metabolic profiles and dynamic process parameters. The first step of proposed method is focused on finding predictive relation between metabolic profiles and process parameter at all time points using OPLS regression. For each time point, the p(corr) obtained from OPLS model is considered as a differential metabogram and is further assessed using principal components analysis (PCA). CONCLUSIONS: Compared to traditional batch modeling, applying proposed methodology on metabolic data from Chinese Hamster Ovary (CHO) antibody production characterized the dynamic relation between metabolic profiles and critical process parameters.

BoT-Net: a lightweight bag of tricks-based neural network for efficient LncRNA-miRNA interaction prediction.

BACKGROUND AND OBJECTIVE: Interactions of long non-coding ribonucleic acids (lncRNAs) with micro-ribonucleic acids (miRNAs) play an essential role in gene regulation, cellular metabolic, and pathological processes. Existing purely sequence based computational approaches lack robustness and efficiency mainly due to the high length variability of lncRNA sequences. Hence, the prime focus of the current study is to find optimal length trade-offs between highly flexible length lncRNA sequences. METHOD: The paper at hand performs in-depth exploration of diverse copy padding, sequence truncation approaches, and presents a novel idea of utilizing only subregions of lncRNA sequences to generate fixed-length lncRNA sequences. Furthermore, it presents a novel bag of tricks-based deep learning approach "Bot-Net" which leverages a single layer long-short-term memory network regularized through DropConnect to capture higher order residue dependencies, pooling to retain most salient features, normalization to prevent exploding and vanishing gradient issues, learning rate decay, and dropout to regularize precise neural network for lncRNA-miRNA interaction prediction. RESULTS: BoT-Net outperforms the state-of-the-art lncRNA-miRNA interaction prediction approach by 2%, 8%, and 4% in terms of accuracy, specificity, and matthews correlation coefficient. Furthermore, a case study analysis indicates that BoT-Net also outperforms state-of-the-art lncRNA-protein interaction predictor on a benchmark dataset by accuracy of 10%, sensitivity of 19%, specificity of 6%, precision of 14%, and matthews correlation coefficient of 26%. CONCLUSION: In the benchmark lncRNA-miRNA interaction prediction dataset, the length of the lncRNA sequence varies from 213 residues to 22,743 residues and in the benchmark lncRNA-protein interaction prediction dataset, lncRNA sequences vary from 15 residues to 1504 residues. For such highly flexible length sequences, fixed length generation using copy padding introduces a significant level of bias which makes a large number of lncRNA sequences very much identical to each other and eventually derail classifier generalizeability. Empirical evaluation reveals that within 50 residues of only the starting region of long lncRNA sequences, a highly informative distribution for lncRNA-miRNA interaction prediction is contained, a crucial finding exploited by the proposed BoT-Net approach to optimize the lncRNA fixed length generation process. AVAILABILITY: BoT-Net web server can be accessed at https://sds_genetic_analysis.opendfki.de/lncmiRNA/.

EL-RMLocNet: An explainable LSTM network for RNA-associated multi-compartment localization prediction.

Subcellular localization of Ribonucleic Acid (RNA) molecules provide significant insights into the functionality of RNAs and helps to explore their association with various diseases. Predominantly developed single-compartment localization predictors (SCLPs) lack to demystify RNA association with diverse biochemical and pathological processes mainly happen through RNA co-localization in multiple compartments. Limited multi-compartment localization predictors (MCLPs) manage to produce decent performance only for target RNA class of particular sub-type. Further, existing computational approaches have limited practical significance and potential to optimize therapeutics due to the poor degree of model explainability. The paper in hand presents an explainable Long Short-Term Memory (LSTM) network "EL-RMLocNet", predictive performance and interpretability of which are optimized using a novel GeneticSeq2Vec statistical representation learning scheme and attention mechanism for accurate multi-compartment localization prediction of different RNAs solely using raw RNA sequences. GeneticSeq2Vec generates optimized statistical vectors of raw RNA sequences by capturing short and long range relations of nucleotide k-mers. Using sequence vectors generated by GeneticSeq2Vec scheme, Long Short Term Memory layers extract most informative features, weighting of which on the basis of discriminative potential for accurate multi-compartment localization prediction is performed using attention layer. Through reverse engineering, weights of statistical feature space are mapped to nucleotide k-mers patterns to make multi-compartment localization prediction decision making transparent and explainable for different RNA classes and species. Empirical evaluation indicates that EL-RMLocNet outperforms state-of-the-art predictor for subcellular localization prediction of 4 different RNA classes by an average accuracy figure of 8% for Homo Sapiens species and 6% for Mus Musculus species. EL-RMLocNet is freely available as a web server at (https://sds_genetic_analysis.opendfki.de/subcellular_loc/).

Transferability of miRNA-technology to bioprocessing: Influence of cultivation mode and media.

The biopharmaceutical industry strives for improvement of their production processes. In recent years, miRNAs have been shown to positively impact the production capacity of recombinant CHO cells, especially with regard to difficult to express proteins. Effective and reliable gene regulation of process relevant target genes by miRNAs is a prerequisite for integrating them into the toolbox of industrial cell engineering strategies. However, most studies rely on transient transfection of miRNA mimics; there is low standardization in evaluation of miRNA function and little knowledge on transferability of effects found during transient expression to stable expression during industry relevant fed-batch cultivation. In order to provide more insight into this topic, we used the pcDNA6.2 vector for stable miRNA overexpression during batch and fed-batch cultivation in CHO DG44 cells, optimized the vector, and compared the miRNA levels and effects with those achieved by transfection of miRNA mimics. We found that miR-1 downregulated TWF1 mRNA in different recombinant CHO DG44 clones in a dose-dependent manner during transient batch cultivation. Cells stably overexpressing miR-1 also showed a TWF1 mRNA downregulation when cultivated in batch mode using in-house medium 1. However, when the cells stably overexpressing miR-1 were cultivated in fed-batch mode using in-house medium 2. Consequently, a change of cultivation mode and medium seems to have an impact on target gene regulation by miRNA. Taken together, our findings highlight the importance to standardize miRNA evaluations and test miRNAs in the final application environment.

MirLocPredictor: A ConvNet-Based Multi-Label MicroRNA Subcellular Localization Predictor by Incorporating k-Mer Positional Information.

MicroRNAs (miRNA) are small noncoding RNA sequences consisting of about 22 nucleotides that are involved in the regulation of almost 60% of mammalian genes. Presently, there are very limited approaches for the visualization of miRNA locations present inside cells to support the elucidation of pathways and mechanisms behind miRNA function, transport, and biogenesis. MIRLocator, a state-of-the-art tool for the prediction of subcellular localization of miRNAs makes use of a sequence-to-sequence model along with pretrained k-mer embeddings. Existing pretrained k-mer embedding generation methodologies focus on the extraction of semantics of k-mers. However, in RNA sequences, positional information of nucleotides is more important because distinct positions of the four nucleotides define the function of an RNA molecule. Considering the importance of the nucleotide position, we propose a novel approach (kmerPR2vec) which is a fusion of positional information of k-mers with randomly initialized neural k-mer embeddings. In contrast to existing k-mer-based representation, the proposed kmerPR2vec representation is much more rich in terms of semantic information and has more discriminative power. Using novel kmerPR2vec representation, we further present an end-to-end system (MirLocPredictor) which couples the discriminative power of kmerPR2vec with Convolutional Neural Networks (CNNs) for miRNA subcellular location prediction. The effectiveness of the proposed kmerPR2vec approach is evaluated with deep learning-based topologies (i.e., Convolutional Neural Networks (CNN) and Recurrent Neural Network (RNN)) and by using 9 different evaluation measures. Analysis of the results reveals that MirLocPredictor outperform state-of-the-art methods with a significant margin of 18% and 19% in terms of precision and recall.

Rerouting of fibroblast growth factor 2 to the classical secretory pathway results in post-translational modifications that block binding to heparan sulfate proteoglycans.

FGF-2 is a proangiogenic growth factor secreted by unconventional means. It is unknown why FGF-2 takes an ER/Golgi-independent secretory route. We find that secretion of FGF-2 via the ER/Golgi system causes post-translational modifications that prevent binding to heparan sulfate proteoglycans (HSPGs), an interaction that is critically important for both FGF-2 storage and signal transduction. This loss of function is due to artificial O-glycosylation mainly resulting in the addition of glycosaminoglycan chains of the chrondroitin sulfate type. Our findings suggest that the unconventional mechanism of FGF-2 export is an ancient pathway of protein secretion that, in the course of evolution, has been kept due to the inability of the classical secretory pathway to export FGF-2 in a functional form.

Recombinase-mediated cassette exchange (RMCE) - a rapidly-expanding toolbox for targeted genomic modifications.

Starting in 1991, the advance of Tyr-recombinases Flp and Cre enabled superior strategies for the predictable insertion of transgenes into compatible target sites of mammalian cells. Early approaches suffered from the reversibility of integration routes and the fact that co-introduction of prokaryotic vector parts triggered uncontrolled heterochromatization. Shortcomings of this kind were overcome when Flp-Recombinase Mediated Cassette Exchange entered the field in 1994. RMCE enables enhanced tag-and-exchange strategies by precisely replacing a genomic target cassette by a compatible donor construct. After "gene swapping" the donor cassette is safely locked in, but can nevertheless be re-mobilized in case other compatible donor cassettes are provided ("serial RMCE"). These features considerably expand the options for systematic, stepwise genome modifications. The first decade was dominated by the systematic generation of cell lines for biotechnological purposes. Based on the reproducible expression capacity of the resulting strains, a comprehensive toolbox emerged to serve a multitude of purposes, which constitute the first part of this review. The concept per se did not, however, provide access to high-producer strains able to outcompete industrial multiple-copy cell lines. This fact gave rise to systematic improvements, among these certain accumulative site-specific integration pathways. The exceptional value of RMCE emerged after its entry into the stem cell field, where it started to contribute to the generation of induced pluripotent stem (iPS-) cells and their subsequent differentiation yielding a variety of cell types for diagnostic and therapeutic purposes. This topic firmly relies on the strategies developed in the first decade and can be seen as the major ambition of the present article. In this context an unanticipated, potent property of serial Flp-RMCE setups concerns the potential to re-open loci that have served to establish the iPS status before the site underwent the obligatory silencing process. Other relevant options relate to the introduction of composite Flp-recognition target sites ("heterospecific FRT-doublets"), into the LTRs of lentiviral vectors. These "twin sites" enhance the safety of iPS re-programming and -differentiation as they enable the subsequent quantitative excision of a transgene, leaving behind a single "FRT-twin". Such a strategy combines the established expression potential of the common retro- and lentiviral systems with options to terminate the process at will. The remaining genomic tag serves to identify and characterize the insertion site with the goal to identify genomic "safe harbors" (GOIs) for re-use. This is enabled by the capacity of "FRT-twins" to accommodate any incoming RMCE-donor cassette with a compatible design.

Recombinase-mediated cassette exchange (RMCE): traditional concepts and current challenges.

Traditional DNA transduction routes used for the modification of cellular genomes are subject to unpredictable alterations, as the cell-intrinsic repair machinery may affect both the integrity of the transgene and the recipient locus. These problems are overcome by recombinase-mediated cassette exchange (RMCE) approaches enabling predictable expression patterns by the nondisruptive insertion of a gene cassette at a pre-characterized genomic locus. The destination is marked by a "tag" consisting of two heterospecific recombination target sites (RTs) at the flanks of a selection marker. Provided on a circular donor vector, an analogous cassette encoding the gene of interest can cleanly replace the resident cassette under the influence of a site-specific recombinase. RMCE was first based on the yeast integrase Flp but had to give way to the originally more active phage-derived Cre enzyme. To be effective, both Tyr-recombinases have to be applied at a considerable concentration, which, in the case of Cre, triggers endonucleolytic activities and therefore cellular toxicity. This review addresses the particularities of both recombination routes depending on the structure of the synaptic complex and on improved integrase and RT variants. While the performance of Flp-RMCE can now firmly rely on optimized Flp variants and multiple sets of functional target sites (FRTs), the Cre system suffers from the promiscuity of its RT mutants, which is explained in molecular terms. At present, RMCE enters applications in the stem cell field. Remarkable efforts are noted in the framework of various mouse mutagenesis programs, which, in their first phase, have targeted virtually all genes and now start to shift their emphasis from gene trapping to gene modification.

Cell-surface heparan sulfate proteoglycans are essential components of the unconventional export machinery of FGF-2.

FGF-2 is an unconventionally secreted lectin that transmits proangiogenic signals through a ternary complex with high-affinity FGF receptors and heparan sulfate proteoglycans (HSPGs). Although FGF-2 signal transduction is understood in great detail, its mechanism of release from cells, which is independent of the classical secretory pathway, remains elusive. To test the hypothesis that FGF-2 secretion is linked to its cell-surface ligands, we studied FGF-2 release using mutants defective for HSPG binding and cells with impaired HSPG biosynthesis. Here, we report that a functional interaction between FGF-2 and HSPGs is required for net export of FGF-2 from mammalian cells. FGF-2 release requires extracellular, membrane-proximal HSPGs. We propose that extracellular HSPGs form a molecular trap that drives FGF-2 translocation across the plasma membrane.

Direct transport across the plasma membrane of mammalian cells of Leishmania HASPB as revealed by a CHO export mutant.

Leishmania HASPB is a lipoprotein that is exported to the extracellular space from both Leishmania parasites and mammalian cells via an unconventional secretory pathway. Exported HASPB remains anchored in the outer leaflet of the plasma membrane mediated by myristate and palmitate residues covalently attached to the N-terminal SH4 domain of HASPB. HASPB targeting to the plasma membrane depends on SH4 acylation that occurs at intracellular membranes. How acylated HASPB is targeted to the plasma membrane and, in particular, the subcellular site of HASPB membrane translocation is unknown. In order to address this issue, we screened for clonal CHO mutants that are incapable of exporting HASPB. A detailed characterization of such a CHO mutant cell line revealed that the expression level of the HASPB reporter molecule is unchanged compared to CHO wild-type cells; that it is both myristoylated and palmitoylated; and that it is mainly localized to the plasma membrane as judged by confocal microscopy and subcellular fractionation. However, based on a quantitative flow cytometry assay and a biochemical biotinylation assay of surface proteins, HASPB transport to the outer leaflet of the plasma membrane is largely reduced in this mutant. From these data, we conclude that the subcellular site of HASPB membrane translocation is the plasma membrane as the reporter molecule accumulates in this location when export is blocked. Thus, these results allow us to define a two-step process of HASPB cell surface biogenesis in which SH4 acylation of HASPB firstly mediates intracellular targeting to the plasma membrane. In a second step, the plasma membrane-resident machinery, which is apparently disrupted in the CHO mutant cell line, mediates membrane translocation of HASPB. Intriguingly, the angiogenic growth factor FGF-2, another protein secreted by unconventional means, is shown to be secreted normally from the HASPB export mutant cell line. These observations demonstrate that the export machinery component defective in the export mutant cell line functions specifically in the HASPB export pathway.

Unconventional protein secretion: membrane translocation of FGF-2 does not require protein unfolding.

Endoplasmic reticulum/Golgi-dependent protein secretion depends on signal peptides that mediate membrane translocation of nascent secretory proteins into the lumen of the endoplasmic reticulum. Classical secretory proteins are transported across the membrane of the endoplasmic reticulum in an unfolded conformation, which is similar to protein import into mitochondria. This process is mediated by Sec61, the protein-conducting channel of the endoplasmic reticulum. Employing both FACS-based in vivo transport assays and confocal microscopy, we now show that fibroblast growth factor 2 (FGF-2), a pro-angiogenic mediator exported from mammalian cells by an unconventional secretory pathway, does not need to be unfolded in order to be released into the extracellular space. These findings suggest that the molecular apparatus mediating export of FGF-2 is not only distinct from classical translocation machineries in terms of molecular identity but also operates in a mechanistically distinct manner that allows membrane translocation of FGF-2 in a folded conformation.

Unconventional secretion of fibroblast growth factor 2 is mediated by direct translocation across the plasma membrane of mammalian cells.

Fibroblast growth factor 2 (FGF-2) is a pro-angiogenic mediator that is secreted by both normal and neoplastic cells. Intriguingly, FGF-2 has been shown to be exported by an endoplasmic reticulum/Golgi-independent pathway; however, the molecular machinery mediating this process has remained elusive. Here we introduce a novel in vitro system that functionally reconstitutes FGF-2 secretion. Based on affinity-purified plasma membrane inside-out vesicles, we demonstrate post-translational membrane translocation of FGF-2 as shown by protease protection experiments. This process is blocked at low temperature but apparently does not appear to be driven by ATP hydrolysis. FGF-2 membrane translocation occurs in a unidirectional fashion requiring both integral and peripheral membrane proteins. These findings provide direct evidence that FGF-2 secretion is based on its direct translocation across the plasma membrane of mammalian cells. When galectin-1 and macrophage migration inhibitory factor, other proteins exported by unconventional means, were analyzed for translocation into plasma membrane inside-out vesicles, galectin-1 was found to be transported as efficiently as FGF-2. By contrast, migration inhibitory factor failed to traverse the membrane of inside-out vesicles. These findings establish the existence of multiple distinct secretory routes that are operational in the absence of a functional endoplasmic reticulum/Golgi system.

Biosynthetic FGF-2 is targeted to non-lipid raft microdomains following translocation to the extracellular surface of CHO cells.

Basic fibroblast growth factor (FGF-2) is a secretory protein that lacks a signal peptide. Consistently, FGF-2 has been shown to be secreted by an ER-Golgi-independent mechanism; however, the machinery mediating this process remains to be established at the molecular level. Here we introduce a novel experimental system based on flow cytometry that allows the quantitative assessment of non-classical FGF-2 secretion in living cells. Stable cell lines have been created by retroviral transduction that express various kinds of FGF-2-GFP fusion proteins in a doxicyclin-dependent manner. Following induction of protein expression, biosynthetic FGF-2-GFP is shown to translocate to the outer surface of the plasma membrane as determined by both fluorescence activated cell sorting (FACS) and confocal microscopy. Both N- and C-terminal GFP tagging of FGF-2 is compatible with FGF-2 export, which is shown to occur in a controlled fashion rather than through unspecific release. The experimental system described has strong implications for the identification of both FGF-2 secretion inhibitors and molecular components involved in FGF-2 secretion. In the second part of this study we made use of the FGF-2 export system described to analyze the fate of biosynthetic FGF-2-GFP following export to the extracellular space. We find that secreted FGF-2 fusion proteins accumulate in large heparan sulfate proteoglycan (HSPG)-containing protein clusters on the extracellular surface of the plasma membrane. These microdomains are shown to be distinct from caveolae-like lipid rafts known to play a role in FGF-2-mediated signal transduction. Since CHO cells lack FGF high-affinity receptors (FGFRs), it can be concluded that FGFRs mediate the targeting of FGF-2 to lipid rafts. Consistently, FGF-2-GFP-secreting CHO cells do not exhibit increased proliferation activity. Externalization and deposition of biosynthetic FGF-2 in HSPG-containing protein clusters are independent processes, as a soluble secreted intermediate was demonstrated. The balance between intracellular FGF-2 and HSPG-bound secreted FGF-2 is shown not to be controlled by the availability of cell surface HSPGs, indicating that the FGF-2 secretion machinery itself is rate-limiting.