PU.1 is a B-cell- and macrophage-specific transcription factor. By an electrophoretic mobility shift assay and dimethyl sulfate methylation interference assays, we show that PU.1 binds to DNA sequences within the immunoglobulin kappa 3' enhancer (kappa E3'). Binding of PU.1 to the kappa E3' enhancer assists the binding of a second tissue-restricted factor, NF-EM5, to an adjacent site. Binding of NF-EM5 to kappa E3' DNA sequences requires protein-protein interaction with PU.1 as well as specific protein-DNA interactions. This is the first known instance of PU.1 interacting with another cellular protein. NF-EM5 does not cofractionate with PU.1, suggesting that it is a distinct protein and is not a posttranslational modification of PU.1. UV-crosslinking studies and elution from sodium dodecyl sulfate-polyacrylamide gels indicate that NF-EM5 is a protein of approximately 46 kDa. Site-directed mutagenesis studies of the PU.1- and EM5-binding sites indicate that these sites play important roles in kappa E3' enhancer activity. By using a series of PU.1 deletion constructs, we have identified a region in PU.1 that is necessary for interaction with NF-EM5. This segment encompasses a 43-amino-acid region with PEST sequence homology, i.e., one that is rich in proline (P), glutamic acid (E), serine (S), and threonine (T).

Damage to neonatal and adult tissues always incites an influx of inflammatory neutrophils and macrophages. Besides clearing the wound of invading microbes, these cells are believed to be crucial coordinators of the repair process, acting both as professional phagocytes to clear wound debris and as a major source of wound growth factor signals. Here we report wound healing studies in the PU.1 null mouse, which is genetically incapable of raising the standard inflammatory response because it lacks macrophages and functioning neutrophils. Contrary to dogma, we show that these "macrophageless" mice are able to repair skin wounds with similar time course to wild-type siblings, and that repair appears scar-free as in the embryo, which also heals wounds without raising an inflammatory response. The growth factor and cytokine profile at the wound site is changed, cell death is reduced, and dying cells are instead engulfed by stand-in phagocytic fibroblasts. We also show that hyperinnervation of the wound site, previously believed to be a consequence of inflammation, is present in the PU.1 null wound, too.

Osteoclasts are multinucleated cells and the principal resorptive cells of bone. Although osteoclasts are of myeloid origin, the role of haematopoietic transcription factors in osteoclastogenesis has not been explored. Here we show that messenger RNA for the myeloid- and B-cell-specific transcription factor PU.1 progressively increases as marrow macrophages assume the osteoclast phenotype in vitro. The association between PU.1 and osteoclast differentiation was confirmed by demonstrating that PU.1 expression increased with the induction of osteoclastogenesis by either 1,25-dihydroxyvitamin D3 or dexamethasone. Consistent with the participation of PU.1 in osteoclastogenesis, we found that the development of both osteoclasts and macrophages is arrested in PU.1-deficient mice. Reflecting the absence of osteoclasts, PU.1-/- mice exhibit the classic hallmarks of osteopetrosis, a family of sclerotic bone diseases. These animals were rescued by marrow transplantation, with complete restoration of osteoclast and macrophage differentiation, verifying that the PU.1 lesion is intrinsic to haematopoietic cells. The absence of both osteoclasts and macrophages in PU.1-mutant animals suggests that the transcription factor regulates the initial stages of myeloid differentiation, and that its absence represents the earliest developmental osteopetrotic mutant yet described.

to 20 min. Utricles were explanted to chilled Medium-199 containing 25 mM Hepes buffer and Hanks' salts (Gibco). The sensory epithelia were isolated, and the otolithic membranes were removed with fine forceps. The culture chambers contained small wells made from cover glasses and Polyallomer rings 9 mm in diameter attached with Silastic adhesive. They were coated with Cell-Tak (Collaborative Research) before one or two utricles were placed in each well with 50 gl of medium. The medium consisted of Medium-199 with Earle's salts, 26 mM sodium bicarbonate, 25 mM Hepes, 0.69 mM L-glutamine (Gibco), supplemented with 20% fetal bovine serum (FBS) (Gibco), penicillin (10 units/mI) and Fungizone (25 ng/mi). Cultures were maintained at 37rC in a 5% CO2 environment. Guinea pigs become sexually mature at 4 to 8 weeks of age [J. E. Wagner and P. J. Manning, The Biology of the Guinea Pig (Academic Press, New York, 1976), p. 9]. Seven of the specimens were older than 8 weeks, and all were at least 6 weeks old. The labeling observed was comparable in utricles from older and younger specimens. All protocols were in accordance with the University of Virginia's guidelines for use of animals in research. 5. Cultures were incubated for 24 hours in Medium-199 with 20% FBS that contained 0.5 to 1.0 mM neomycin or 1.0 to 2.0 mM gentamicin. 6. After 24 hours, cultures were rinsed twice with Medium-199, and fresh medium without any aminoglycoside antibiotics was added. The aminoglycoside-free media contained either of two mitotic tracers: [3H]methyl-thymidine (0.8 gCi/ml, 65Ci/mmol) or 5-bromo-2'-deoxyuridine (BrdU, 3 gg/ml) in solution with 5-fluoro-2'-deoxyuridine (0.27 gg/ml) from Amersham.