(A) Thymus cDNA from three different +/+ or thu/thu mice, PCR amplified with primers in Ptprc exons 3 and 7. Specific RNA splice products were deduced on the basis of their sequence-predicted sizes: RO, 71 bp; RB, 218 bp; RAB, 347 bp; RBC, 359 bp; RABC, 488 bp.
(B) Ptprc exon 4, 5, and 6 inclusion quantified in different cell subsets by geometric-mean fluorescence staining with antibodies against the corresponding CD45RA, CD45RB, and CD45RC segments. Relative CD45 expression was measured in parallel with antibody to the CD45.2 epitope in a membrane-proximal domain unaffected by alternative splicing. Abbreviations are as follows: DN, CD4⁻8⁻ double-negative thymocytes; DP, CD4⁺8⁺ double-positive thymocytes; CD4 SP, CD4⁺ single-positive thymocytes; CD8 SP, CD8⁺ single-positive thymocytes; B, spleen B cells; CD4, spleen CD4 T cells; CD8, spleen CD8 T cells. Data are from six animals of each genotype. Columns indicate means, and error bars indicate standard deviations from n = 6 mice.

(A) Ribbon representation of the hnRNPLL RRM1 domain (residues 124–202), including side chains at the site of mutation (V136, yellow) and RNA contact residues conserved in other RRM domains (His128, Gln162). Red spheres: backbone amides broadened beyond detection in the NMR spectra of the V136D mutant. Orange spheres: backbone amides for which the mutation results in different chemical shifts.
(B) Same as (A) except that spheres identify backbone amides with different chemical shifts in the presence of the 15-mer RNA 5′-CCUUACCUGCACGCA-3′, with red colors indicating amide protons broadened beyond detection in the presence of the 15-mer RNA.
(C and D) Same as (A) and (B), but with ribbon representation of the PTB1 RRM1 domain (PDB: 1SJQ).
(E) Binding affinity of RRM1 domain to 15-mer RNA, measured by NMR spectroscopy at 25°C. K_d values obtained from the chemical-shift changes of His128, Asn169 and Ala192 were 0.1 ± 0.5 μM, 0.7 ± 0.7 μM, and 1.5 ± 1.3 μM, respectively, indicating an upper limit of about 1.5 μM.
(F) Same as (E), but for V136D mutant. Corresponding K_d values were 1.1 ± 0.4 μM, 1.0 ± 0.4 μM, and 2.0 ± 0.7 μM, indicating an upper limit of about 2 μM.
(G) Binding affinity measured by isothermal calorimetry at 15°C. The continuous line represents the least-squares fit of the data, yielding a K_d value of 1.9 ± 0.4 μM.
(H) Same as (G), but for the V136D mutant. The fitted K_d value is 3.1 ± 1.4 μM.
(I) DSC curves of mutant (red) or WT (black) RRM1 domains. Both proteins precipitated irreversibly at high temperature.

Summary of exon-array analysis of mRNA from naive and memory CD4⁺ and CD8⁺ T cells and SP thymocytes sorted from thu/thu and +/+ mice (four biological replicates of each cell and genotype). Of the 228 most highly ranked Hnrpll-regulated genes, those assigned to Group A (red font) and Group B (blue font) contained Hnrpll-dependent probe sets corresponding to known alternatively spliced exons in reference mouse transcripts. Asterisks denote genes with transcript or published data for alternative splicing of the same exons in man. Chevrons denote genes with published data for alternative splicing of the same exons in neural tissue. Groups A and C (green font): genes with Hnrpll-dependent exons exhibiting consistently increased signal in thu/thu. Groups B and D (purple font): genes with Hnrpll-dependent exons exhibiting consistently decreased signal in thu/thu. Black font denotes genes in Group E, with Hnrpll-dependent probe sets that appear to reflect alternative promoter use, and in Group F, with a complex pattern of changes in Hnrpll-dependent probe sets.

Flow-cytometric analysis of cells from WT +/+ (unfilled histograms and symbols), thu/+ (gray) and thu/thu (filled black) mice. This scheme of symbols and colors applies to all panels.
(A) CD44 on CD4⁺ or CD8⁺ blood T cells. The percentage of memory (CD44^hi) and naive (CD44^lo) subsets is shown.
(B) Staining for specific CD45 isoforms or an epitope common to all isoforms (CD45.2), gated on splenic CD4⁺ or CD8⁺ T cells or B cells.
(C) Number of CD4⁺ or CD8⁺ SP T cells in the thymus at 6 weeks.
(D) Number of naive or memory CD4⁺ and CD8⁺ T cells in the spleen at different ages.
(E) T cells in irradiated mice reconstituted with an equal mixture of +/+ CD45.1 and thu/thu CD45.2 (CD45.1-negative) bone marrow. Control mixed chimeras were reconstituted +/+ CD45.1 and +/+ CD45.2 bone marrow. Left panel shows percentage of CD4⁺ or CD8⁺ T cells that are CD45.1-negative, representing thu/thu T cells in the test mixed chimeras (filled squares) or +/+ T cells in the control mixed chimeras (open circles). Right panel shows percentage of CD4⁺ or CD8⁺ T cells that are CD45.1-positive, representing +/+ T cells either in the test mixed chimeras (filled squares) or in the control chimeras (open circles).
(F) Percentage of memory T cells among the CD45.1-negative T cells in (E), representing thu/thu T cells (filled squares) or +/+ T cells (open circles).
(G) A 50:50 mixture of +/+ CD45.1 and thu/thu CD45.2 thymocytes was injected into Rag1^−/−° mice. The percentage of T cells of each donor type is shown at the time of injection and in the blood at different times after transfer.

(A) Meiotic mapping of thunder to a 2.1 Mb interval derived from the B6-thunder strain.
(B) Candidate RNA-regulating genes within the interval.
(C) T→A substitution in thunder mRNA nucleotide 684, codon 136.
(D) Predicted domains of the hnRNPLL and hnRNPL proteins. Abbreviations are as follows: RRM, RNA recognition motif domain; PR, proline rich domain.
(E) Amino acid sequence alignment of the hnRNPLL RRM1 domain from different vertebrate species and with corresponding RRM domains from hnRNPL and PTB1. Gold boxes denote two RNP motifs that form a central binding site for RNA, with key RNA contacts on H and Q residues. The regular secondary-structure elements in the RRM1 domain of hnRNPLL are indicated at the top.
(F) thu/thu T lymphoblasts were transduced with a bicistronic retroviral vector containing WT Hnrpll cDNA and EGFP or containing WT EGFP alone. Flow-cytometric staining for CD45RC is shown gated on GFP⁺ cells (gray histograms) compared with nontransduced GFP° cells in the same culture (open histograms), showing correction of CD45RC silencing by the WT cDNA.
(G) Relative mRNA abundance of Hnrpll (upper panel) and Hnrpl (lower panel) in WT naive and memory CD4⁺ and CD8⁺ T cells measured by qRT-PCR, normalized to Ube2D1. Columns indicate means, and error bars indicate standard deviations from n = 3 mice.

Log2 mean and standard error of the signal for each probe set is shown for either CD4⁺ or CD8⁺ T cells (memory thu/thu in red, memory +/+ in green, naive thu/thu in blue, naive +/+ in purple), normalized against the mean signal on the same probe set in corresponding naive +/+ cells. Grey arrows indicate sites of transcript initiation.
(A and B) Examples of genes in Group A, in which signal is consistently increased in thu/thu cells for probe sets corresponding to known alternatively spliced internal exons.
(C and D) Examples of genes in Group B, in which signal is consistently decreased in thu/thu cells for probe sets corresponding to known alternatively spliced internal exons.
(E and F) Examples of genes in which multiple internal but noncontiguous exons, corresponding to known transcript isoforms, show consistent Hnrpll-dependent changes.
(G and H) Examples of genes in which variable silencing of an internal splice donor creates alternative 3′ mRNA and carboxy termini.