To the Editor:
In the December 15, 1997 issue of BLOOD, Miraglia et al1reported the sequence of the AC133 antigen, a novel 865 amino acid–residue surface glycoprotein found in human hematopoietic stem and progenitor cells. In the November 11, 1997 issue of Proceedings of the National Academy of Sciences of the United States of America, our group2 described the novel protein prominin, an 858 amino acid–residue polytopic membrane protein specific to microvilli on the apical surface of various murine embryonic and adult epithelia. We wish to bring to the attention of the readers of BLOOD that the human AC133 antigen and mouse prominin are highly related proteins that share the same 5-transmembrane topology and show an average 60% amino acid–sequence identity (Fig1) and to discuss implications of this relationship.
First, the AC133 antigen may be the human homologue of mouse prominin. This would imply that (1) there is considerable species variation between human and murine prominin/AC133 antigen, in particular in the extracellular domains and (2) the cellular distribution of prominin/AC133 antigen is broader than assumed from the results of the two initial studies.1-3 Evidence consistent with this possibility includes the following points. The monoclonal antibody 13A4 used to isolate the prominin cDNA clone from adult mouse kidney recognizes a similar, if not identical, protein in the neuroepithelium of mouse embryos, and EST sequences from human retina, which originates from the neuroepithelium, are more closely related to the human AC133 sequence than to that of mouse kidney prominin. Also, AC133 immunoreactivity is detected in the human NT2 cell line,3which exhibits neuroepithelial features. If the AC133 antigen is the human homologue of mouse prominin, the lack of AC133 immunoreactivity in human kidney may reflect the fact that the monoclonal antibody AC133 recognizes a glycosylated structure,1 which would imply differential glycosylation of prominin/AC133 antigen in various cell types.
Second, prominin and the AC133 antigen may be distinct members of a novel family of membrane proteins. If so, the difference between epithelial and nonepithelial members of this protein family may account for some, if not most, of the sequence variation between mouse kidney prominin and the human hematopoietic stem cell antigen AC133. Evidence consistent with the existence of multiple members of the prominin/AC133 antigen family within a given species includes the following examples. Northern blot analysis using the mouse kidney prominin cDNA as probe reveals the presence of similar RNAs in mouse kidney and gut, whereas immunoreactivity is detected in kidney, but not gut, using either the 13A4 monoclonal antibody2 or two polyclonal antibodies raised against recombinant mouse kidney prominin (D. Corbeil, unpublished data). Likewise, AC133 transcripts were detected in various human tissues whereas AC133 immunoreactivity was confined to human bone marrow.1 Interestingly, theCaenorhabditis elegans genome4contains, in addition to the predicted protein F08B12.1 that is strikingly similar to mouse prominin2 and the human AC133 antigen (Fig 1), at least two other open reading frames (M28.9 and M28.8, accession no. Z49911) that predict proteins related in size and hydropathy pattern to mouse prominin and the human AC133 antigen.
It will be important to determine whether a key feature of prominin, its selective targetting to plasma membrane protrusions, is also observed for the AC133 antigen in hematopoietic stem cells. If so, and on the assumption that for any given species there will be more than one prominin/AC133 protein, the prominin of epithelial cells could be referred to as E-prominin and the AC133 antigen perhaps as H-prominin (for hematopoietic stem cell prominin).
Even more importantly, the function of prominin/AC133 proteins remains to be established. One may speculate whether the expression of prominin in various murine embryonic epithelia and of the AC133 antigen in human hematopoietic stem cells provides a clue for a role of these novel membrane proteins in cell determination.