washed with Ca2+, Mg2+-containing 1× phosphate
buffered saline (PBS) and blocked with 3% fetal bovine serum (FBS) in PBS for 30
min at 4°C. Cells were aliquoted (100 μl/tube) for binding and
staining with 2.5 μg/ml of FITC- or Alexa Fluor (AF)-conjugated
antibodies. Sorting was performed using a BD Biosciences fluorescence-activated
cell sorting (FACS) DiVa High-Speed Cell Sorter (San Diego, CA) with 350, 488,
and 633 nm lasers. Anti-CD34 to Alexa Fluor 488 and anti-Stro-1 were conjugated
to Alexa Fluor 488, anti-CD45 to FITC, anti-CD90 to PE, anti-CD105 to Alexa
Fluor 488, and to the appropriate isotype controls (all from BD Pharmingen Inc.,
San Diego, CA except anti-Stro-1, which was obtained from Invitrogen, Carlsbad,
CA). The data were analyzed using FlowJo software (Tree Star, San Carlos, CA).
Propidium iodide was used to exclude dead cells, and percentages of positively
stained-cells were calculated by subtracting the value of isotype controls.
Cells were negatively selected for CD34 (NOVUS, catalog# NB600–107 AF
488) and CD45 (BD Pharmingen, Catalog# 554877 FITC) and positively selected for
CD90 (BD Pharmingen Catalog# 554898 PE), CD105 (Application.Inc, catalog#
CA1725), and Stro-1 (Invitrogen, catalog# 398407).