Three recently published works represent the first descriptions of a potential new family of 5-trans-membrane cell surface glycoproteins.1,2,3 The human AC133 molecule2,3may indeed be the homologue of mouse prominin,1,4 and Drs Weigmann and colleagues have very carefully discussed the evidence for and against this relationship. We agree with the arguments put forward by these authors, and that further work is clearly needed to examine the possible localization of AC133 antigen to plasma membrane protrusions, and to assess the expression of prominin on murine hematopoietic stem cells.
If prominin proves to be a marker of primitive murine stem cells, as AC133 is in humans, its value in studies of murine hematopoiesis would be considerable. To assist these studies, we have also isolated and characterized a prominin-like cDNA from mouse brain. This murine sequence differs slightly from the prominin sequence, as we have identified an additional block of nine amino acids within the N-terminal domain. These nine amino acids are strongly homologous to the corresponding human sequence (Fig 1).The cDNA isolated from mouse brain therefore predicts a protein of 867 amino acids, with a molecular weight of 97.3 kD.5 The presence of murine AC133 transcript in a variety of mouse tissues was assessed by Northern analysis (Fig 2). A 4.4-k mRNA transcript was detectable in mouse embryo at 7 days, peaking at 11 days, and tapering off at days 15 and 17 while still being clearly detectable. In adult mouse tissue, the transcript is weakly detectable in brain, lung, skeletal muscle, and more intensely in kidney and testis. Molecular analysis of the three proteins from human, mouse and Caenorhabtidis elegans6 indicate that all three contain leucine zipper consensus sequences (on the first extracellular loop in the human and mouse, and on the second in C elegans). There are conserved cysteine residues in all three molecules appearing at the beginning of each small cytoplasmic loop. Although the mouse and human proteins exhibit strong conservation in all five predicted transmembrane sequences, the greatest degree of conservation with the C elegans sequence is within TM3, where 7 out of 22 amino acids are conserved in all three species. There are seven tyrosine residues within the C-terminal cytoplasmic tail in C elegans;however, none of them occur within a tyrosine phosphorylation consensus sequence. The cytoplasmic tails of the mouse and human proteins contain five tyrosine residues, the first of which is a tyrosine phosphorylation consensus sequence that is completely identical in both proteins. Another sequence7 in the Expressed Sequence Tag database, derived from murine embryo, has approximately 25% homology to the human and mouse AC133 antigen/prominin sequences, indicating that these two molecules may indeed be the first described of a new family of cell surface receptors.
The expression of such a gene or gene family on hematopoeitic human stem cells, mouse embryo, and adult tissues, and in a phylogenetically less evolved organism such as C elegans may indicate an important role for this five-transmembrane protein structure in development and/or cell signaling. In particular, the localization of the prominin molecule to cellular protrusions places this molecule in an optimal locale to participate in cellular communications via cell-cell contact, and possibly signaling.