Reconstitution of the minor H antigen HA-3 with HPLC-fractionated peptides extracted from HLA-A*0101 molecules. HLA-A*0101–associated peptides were purified from 5 × 10¹⁰ Rp EBV-BLCLs and fractionated by RP-HPLC as described in “Patient, materials, and methods.” Aliquots of each fraction (corresponding to 4.5 × 10⁹, 4 × 10⁹, and 8 × 10⁹ cell equivalents for each of the 3 respective dimensions of epitope reconstitution shown in panels A-C) were preincubated with ⁵¹Cr-labeled HoDo cells and tested for their ability to reconstitute epitope activity for the HA-3 CTL clone 5Ho11. An E/T ratio of 17:1 was used. (A) First-dimension separation of extracted peptides was achieved using TFA as the ion-pairing agent. The peak in fractions 18 and 19 is biologically active, while the peak in fractions 3 to 6 is due to the acetic acid in the peptide extract sample. (B) Fractions 18 and 19 from panel A were pooled and rechromatographed using HFBA as the ion-pairing agent. (C) Fractions 25 and 26 from panel B were pooled and rechromatographed on a microcapillary column using TFA as the ion-pairing agent. (D) Determination of candidate peptides via mass spectrometry correlated with ⁵¹Cr-release assay. Fractions 46 to 48 were pooled and chromatographed using nanoflow effluent splitter technology, and aliquots of each splitter fraction corresponding to 8 × 10⁹ cell equivalents were incubated with ⁵¹Cr-labeled HoDo target cells and tested for their ability to reconstitute epitope activity as described in “Patient, materials, and methods.” Ion abundances of candidate masses within the MS scan window 155-215 were plotted and correlated to the percent specific ⁵¹Cr release in that same region. Background lysis of HoDo by the CTLs in the absence of any peptides was 13% in panelA, -3% in panel B, 5% in panel C, and -2% in panel D. Positive control lysis was 70% in panel A, 51% in panel B, 88% in panel C, and 60% in panel D.