Production of cyclic peptides and proteins in vivo

Combinatorial libraries of synthetic and natural products are an important source of molecular information for the interrogation of biological targets. Methods for the intracellular production of libraries of small, stable molecules would be a valuable addition to existing library technologies by combining the discovery potential inherent in small molecules with the large library sizes that can be realized by intracellular methods. We have explored the use of split inteins (internal proteins) for the intracellular catalysis of peptide backbone cyclization as a method for generating proteins and small peptides that are stabilized against cellular catabolism. The DnaE split intein from Synechocystis sp. PCC6803 was used to cyclize the Escherichia coli enzyme dihydrofolate reductase and to produce the cyclic, eight-amino acid tyrosinase inhibitor pseudostellarin F in bacteria. Cyclic dihydrofolate reductase displayed improved in vitro thermostability, and pseudostellarin F production was readily apparent in vivo through its inhibition of melanin production catalyzed by recombinant Streptomyces antibioticus tyrosinase. The ability to generate and screen for backbone cyclic products in vivo is an important milestone toward the goal of generating intracellular cyclic peptide and protein libraries.

Production of zebrafish germ-line chimeras from embryo cell cultures

Although the zebrafish possesses many characteristics that make it a valuable model for genetic studies of vertebrate development, one deficiency of this model system is the absence of methods for cell-mediated gene transfer and targeted gene inactivation. In mice, embryonic stem cell cultures are routinely used for gene transfer and provide the advantage of in vitro selection for rare events such as homologous recombination and targeted mutation. Transgenic animals possessing a mutated copy of the targeted gene are generated when the selected cells contribute to the germ line of a chimeric embryo. Although zebrafish embryo cell cultures that exhibit characteristics of embryonic stem cells have been described, successful contribution of the cells to the germ-cell lineage of a host embryo has not been reported. In this study, we demonstrate that short-term zebrafish embryo cell cultures maintained in the presence of cells from a rainbow trout spleen cell line (RTS34st) are able to produce germ-line chimeras when introduced into a host embryo. Messenger RNA encoding the primordial germ-cell marker, vasa, was present for more than 30 days in embryo cells cocultured with RTS34st cells or their conditioned medium and disappeared by 5 days in the absence of the spleen cells. The RTS34st cells also inhibited melanocyte and neuronal cell differentiation in the embryo cell cultures. These results suggest that the RTS34st splenic–stromal cell line will be a valuable tool in the development of a cell-based gene transfer approach to targeted gene inactivation in zebrafish.

Training and search methods for speech recognition.

Speech recognition involves three processes: extraction of acoustic indices from the speech signal, estimation of the probability that the observed index string was caused by a hypothesized utterance segment, and determination of the recognized utterance via a search among hypothesized alternatives. This paper is not concerned with the first process. Estimation of the probability of an index string involves a model of index production by any given utterance segment (e.g., a word). Hidden Markov models (HMMs) are used for this purpose [Makhoul, J. & Schwartz, R. (1995) Proc. Natl. Acad. Sci. USA 92, 9956-9963]. Their parameters are state transition probabilities and output probability distributions associated with the transitions. The Baum algorithm that obtains the values of these parameters from speech data via their successive reestimation will be described in this paper. The recognizer wishes to find the most probable utterance that could have caused the observed acoustic index string. That probability is the product of two factors: the probability that the utterance will produce the string and the probability that the speaker will wish to produce the utterance (the language model probability). Even if the vocabulary size is moderate, it is impossible to search for the utterance exhaustively. One practical algorithm is described [Viterbi, A. J. (1967) IEEE Trans. Inf. Theory IT-13, 260-267] that, given the index string, has a high likelihood of finding the most probable utterance.