Gene probes to detect cross-culture contamination in hormone producing cell lines

Summary Cross-culture contamination of cell lines propagated in continuous culture is a frequent event and particularly difficult to resolve in cells expressing similar phenotypes. We demonstrate that DNA-DNA hybridization to blotted endonuclease-digested cell DNA effectively detects cross-culture contamination to monitor inter-species as well as intra-species cross contamination. An insulin-producing cell-line, Clone-16, originally cloned from a human fetal endocrine pancreatic cell line did not produce human c-peptide as anticipated. DNA from these cells showed no hybridization to the human ALU sequence probe, BLUR, and lacked restriction fragment length polymorphism typical for the human HLA-DQ β-chain gene. Although a human insulin gene probe showed a weak, nonhuman hybridization pattern, a cDNA probe for the Syrian hamster insulin gene hybridized strongly consistent with a single copy hamster insulin gene. Karyotyping confirmed the absence of human chromosomes in the Clone-16 cells while sizes, centromere indices, and banding patterns were identical to Syrian hamster fibroblasts. We conclude that the insulin-producing Clone-16 cells are of Syrian hamster origin and demonstrate the effective use of gene probes to control the origin of cell cultures. This paper is dedicated to the late Lis Lyngsie in much appreciation of her contributions to this study. This work was supported in part by the National Institutes of Health, Bethesda, MD (grants DK 26190 and 33873). I. Matsuba and B. Michelsen were supported by research fellowships from the Juvenile Diabetes Foundation International, and J. Scholler from the Danish Medical Research Council (J.no. 12-5758).     

Isolation of cloned Syrian hamster insulinoma cell lines with limited capacity for insulin production.

Cell lines derived from a Syrian hamster insulinoma have been cloned in agar and maintained continuously in culture in vitro for 1 1/2 years. The cell lines have average doubling times of 48 h, they have modal chromosome numbers between 38 and 39, and they retain the ability to form tumors when injected into Syrian hamsters. The cells do not produce immunoreactive insulin when grown in vitro (less than 0.5 ng/mg protein), but do produce immunoreactive insulin when grown in vivo as tumors (mean value from six determinations = 7.1 +/- 1.5 ng/mg protein).     

A novel human insulinoma-associated cDNA, IA-1, encodes a protein with "zinc-finger" DNA-binding motifs.

A subtraction library was constructed from human insulinoma (beta cell tumor) and glucagonoma (alpha cell tumor) cDNA phagemid libraries. Differential screening of 153 clones with end-labeled mRNAs from insulinoma, glucagonoma, and HeLa cells resulted in the isolation of a novel cDNA clone designated IA-1. This cDNA clone has a 2838-base pair sequence consisting of an open reading frame of 1530 nucleotides, which translates into a protein of 510 amino acids with a pI value of 9.1 and a molecular mass of 52,923 daltons. At the 3'-untranslated region there are seven ATTTA sequences between two polyadenylation signals (AATAAA). The IA-1 protein can be divided into two domains based upon the features of its amino acid sequence. The NH2-terminal domain of the deduced protein sequence (amino acids 1-250) has four classical pro-hormone dibasic conversion sites and an amidation signal sequence, Pro-Gly-Lys-Arg. The COOH-terminal domain (amino acids 251-510) contains five putative "zinc-finger" DNA-binding motifs of the form X3-Cys-X2-4-Cys-X12-His-X3-4-His-X4 which has been described as a consensus sequence for members of the Cys2-His2 DNA-binding protein class. Northern blot analysis revealed IA-1 mRNA in five of five human insulinoma and three of three murine insulinoma cell lines. Expression of this gene was undetectable in normal tissues. Additional tissue studies revealed that the message is expressed in several tumor cell lines of neuroendocrine origin including pheochromocytoma, medullary thyroid carcinoma, insulinoma, pituitary tumor, and small cell lung carcinoma. The restricted tissue distribution and unique sequence motifs suggest that this novel cDNA clone may encode a protein associated with the transformation of neuroendocrine cells.     

The human CCG1 gene, essential for progression of the G1 phase, encodes a 210-kilodalton nuclear DNA-binding protein.

The human CCG1 gene complements tsBN462, a temperature-sensitive G1 mutant of the BHK21 cell line. The previously cloned cDNA turned out to be a truncated form of the actual CCG1 cDNA. The newly cloned CCG1 cDNA was 6.0 kb and encoded a protein with a molecular mass of 210 kDa. Using an antibody to a predicted peptide from the CCG1 protein, a protein with a molecular mass of over 200 kDa was identified in human, monkey, and hamster cell lines. In the newly defined C-terminal region, an acidic domain was found. It contained four consensus target sequences for casein kinase II and was phosphorylated by this enzyme in vitro. However, this C-terminal region was not required to complement tsBN462 mutation since the region encoding the C-terminal part was frequently missing in complemented clones derived by DNA-mediated gene transfer. CCG1 contains a sequence similar to the putative DNA-binding domain of HMG1 in addition to the previously detected amino acid sequences common in nuclear proteins, such as a proline cluster and a nuclear translocation signal. Consistent with these predictions, CCG1 was present in nuclei, possessed DNA-binding activity, and was eluted with similar concentrations of salt, 0.3 to 0.4 M NaCl either from isolated nuclei or from a DNA-cellulose column.     

Systematic binding analysis of the insulin gene transcription control region: insulin and immunoglobulin enhancers utilize similar transactivators.

The 5' regulatory region (-345 to +1) of the rat insulin I gene (Ins-I) was examined for binding to cellular factors with short oligodeoxynucleotide probes. Over 40 binding species were detected. The binding profiles were specific for each cell type studied. We characterized the factors binding two elements crucial for enhancer activity (the Nir and Far boxes) which bear sequence similarity to the microE1, microE2, and microE3 elements of the immunoglobulin heavy-chain enhancer. The Nir box binds three cellular factors that display preferential affinities for microE1, microE2, or microE3, and the Far box binds two factors related to microE2 or microE3. The insulin gene enhancer was mutated at the Nir box element to reflect the sequences of microE1, microE2, or microE3. Ins-microE2 was fully active, Ins-microE3 was partially active, and Ins-microE1 was inactive. Thus, factors similar or identical to nuclear factor NF-microE1, NF-microE2, or NF-microE3 may play a role in the activity of the insulin gene enhancer.     

Coding sequence and growth regulation of the human vimentin gene.

We have established the complete coding sequence of the human vimentin gene. It had 91% homology to the coding sequence of the Syrian hamster vimentin gene (Quax et al., Cell 35:215-223, 1983) and partial homology to several other sequences coding for intermediate filament proteins. The most striking difference between the Syrian hamster and human vimentin genes was in the 3' untranslated region, which was considerably longer in the Syrian hamster. Using RNA blots and a human vimentin cDNA clone from an Okayama-Berg library, we have established that expression of the vimentin gene was growth regulated. The steady-state levels of cytoplasmic vimentin mRNA in 3T3 cells were increased by serum and platelet-derived growth factor, but not by epidermal growth factor, insulin, or platelet-poor plasma. The increase in expression of the vimentin gene that occurred when G0-phase cells were stimulated to proliferate was detected in six different cell types from four different species. The expression of the vimentin gene was also increased when HL60 cells were induced to differentiate by phorbol esters; it decreased when differentiation was induced by retinoic acid.     

Characterization of the single-strand-specific BPV-1 origin binding protein, SPSF I, as the HeLa Pur alpha factor.

SPSF I and II are two cellular proteins which bind specifically to single-stranded DNA. SPSF I and II binding sites are found in the minimal origin of replication of BPV-1 DNA and near the P2 promoter of the cellular c-myc gene. DNA-binding properties of the two proteins to single-stranded oligonucleotides of different lengths and sequences were quantified by determination of DNA-binding constants. The binding constant of SPSF proteins to the lower strand of the BPV-1 origin was determined to be 1.5 x 10(-10) M-1. Peptide sequences derived from purified SPSF I and II revealed the identity of at least one of the SPSF proteins with the so-called HeLa Pur alpha factor. The HeLa Pur alpha factor was identified previously by virtue of its capacity to bind to purine-rich strands of the PUR element found in initiation zones of DNA replication [Bergemann, A.D., Ma,Z.-W. and Johnson, E.M. (1992) Mol. Cell. Biol. 12, 5673-5682]. Expression of the Pur cDNA confirmed the identity of the Pur alpha protein with the 42 kDa SPSF I protein. Analysis of several Pur alpha cDNA clones revealed the existence of an extended 3'-untranslated region in all Pur mRNAs.