Fringe modulation of Jagged1-induced Notch signaling requires the action of beta 4galactosyltransferase-1 -- Data Supplement - Supporting Information -- Proceedings of the National Academy of Sciences

Chen et al. (November 13, 2001) Proc. Natl. Acad. Sci. USA, 10.1073/pnas.241398098.

Supporting Information
Endogenous Fringe in Chinese Hamster Ovary (CHO) Cells.
We performed single tube reverse-transcription (RT)-PCR reactions (Qiagen) to detect endogenous CHO cell fringe transcripts by using total CHO cell RNA with an upstream primer against mouse Lfng codons 194-202 (5’-TGGTTCTGCCACGTGGATGATGACAAC-3’) and one of two degenerate downstream primers against mouse Lfng codons 328-334 (no. 1, 5’-VCCRTAGCTSARGGTVACCTG-3’; no. 2, 5’-VCCATAGCTCAGGGTVACCTG-3’) that span 3 introns in mouse and human genomic DNA. The upstream primer is 100% conserved in mouse lunatic (Lfng), radical (Rfng), and manic (Mfng) fringes, and 93% identical to the equivalent region in the three human fringe-coding sequences. The downstream primers are 90% identical between mouse and human Lfng, Mfng, and Rfng. A single prominent band of the predicted cDNA size for each fringe (approximately 400 nucleotides) was obtained from both reactions.

Sequencing of both reverse transcription (RT)-PCR products revealed the same mixture of two homologous sequences of approximately equal intensity, reflecting the presence of both Rfng and Lfng cDNAs. Over a span of 103 nucleotides (from mouse Lfng codons 207-241), 25 of 27 mouse Rfng-specific nucleotides, 14 of 19 mouse Lfng-specific nucleotides, and 0 of 23 mouse Mfng-specific nucleotides were present in the CHO RT-PCR products. Small sequence differences between hamster and mouse Lfng and Rfng were mainly at the wobble position and reflect species variation at the nucleotide but not the amino acid level. Of 5 human and mouse Lfng-specific nucleotides in the 103-nt span, 4 were present in CHO; of 6 human and mouse Rfng-specific nucleotides, 6 were present in CHO; and of 9 human and mouse Mfng-specific nucleotides, none were present in CHO. Six clones derived from RT-PCR products were sequenced also. Four clones had 93% nucleotide and 99% amino acid identity with mouse Lfng, and two clones had 92% nucleotide and 97% amino acid identity with mouse Rfng. The combined data show that our CHO cells express Lfng and Rfng but not Mfng transcripts. The same conclusions were reported by Shimizu et al. (1) who detected low levels of Rfng and Lfng but no Mfng transcripts by Northern analyses of CHO-r cells (1). Fringe proteins were not detected by Western analyses using commercial Abs (data not shown). However, fringe is active at a low level in CHO cells because elongated O-fucose is present on endogenous Notch (2). This endogenous activity apparently does not interfere with the interpretation of our signaling assay, because transfection of Mfng or Lfng into CHO cells causes an increase in O-fucose elongation on Notch EGF repeats and a change in Notch signaling.

O
-Fucose Glycans from Notch1 EGF1-36 Expressed in Different CHO Glycosylation Mutants. To show that fringe made the predicted O-fucose glycans on Notch extracellular domain in the different CHO glycosylation mutants, O-fucose glycans on a Notch1 construct containing EGF1-36 were prepared from Lec1, Lec2, Lec8, and Lec20 CHO cells stably expressing Mfng. Lec1 cells do not synthesize complex or hybrid N-glycans and consequently have reduced fucose incorporated into N-glycans, whereas Lec2, Lec8, and Lec20 cells have normal incorporation of fucose into N-glycans. In the latter mutant lines, analysis of O-fucose glycans was complicated by the presence of [³H]fucose-labeled N-glycans released with O-glycans during the b -elimination step. Fucose in O-glycans was distinguished from fucose in N-glycans by the fact that only O-fucose becomes reduced to fucitol during b -elimination. All glycans released from Notch1 EGF1-36 by b -elimination were analyzed for the presence of fucose or fucitol by high pH anion-exchange chromatography. The tetrasaccharide peak from Lec1.Mfng cells contained fucitol, demonstrating that it was derived from O-fucose. Peaks that migrated at or near the position of authentic tetrasaccharide from Lec2.Mfng, Lec8.Mfng, or Lec20.Mfng cells contained fucose but no fucitol, showing that these peaks were derived from N-glycans and not from O-fucose glycans. Therefore, these cell lines did not produce detectable amounts of O-fucose tetrasaccharide. Lec2.Mfng cells made O-fucose trisaccharide, whereas both Lec8.Mfng and Lec20.Mfng cells produced only O-fucose disaccharide. Analysis of mouse Notch1 EGF19-23 produced from the different cell lines by the same strategy is shown in the article. In that case, the profiles of O-fucose glycans were not complicated by N-glycans because Notch fragment EGF19-23 contains only a single potential N-glycan site, and N-glycanase treatment effectively removed N-glycans in this case.

References

1. Shimizu, K., Chiba, S., Saito, T., Kumano, K., Takahashi, T. & Hirai, H. (2001) J. Biol. Chem. 276, 25753–25758.

2. Moloney, D. J., Shair, L. H., Lu, F. M., Xia, J., Locke, R., Matta, K. L. & Haltiwanger, R. S. (2000) J. Biol. Chem. 275, 9604–9611.