Analysis of LARD Expression.

A full-length LARD cDNA was ³²P-labeled using a random primer kit (Primeit II, Stratagene). This probe was hybridized to multiple human-tissue Northern blots according to the manufacturer’s instructions (MTN1-MTN2, Stratagene). RNase protection analysis was carried out according to standard protocols (12). A 266-bp complementary probe, covering the 3′ end of the LARD untranslated region, that will hybridize to all LARD isoforms was generated from an SP6 promoter and hybridized with 10 μg of total cellular RNA.

Alternative Splicing.

Peripheral blood lymphocytes (PBL) were prepared by Ficoll density gradient centrifugation of cells prepared from the buffy coat of 1 unit of human blood. CD4⁺, CD8⁺, and B cells were purified by absorption onto magnetic beads coated with antibodies to CD4, CD8, and CD19, respectively. Macrophages were prepared by adhesion to plastic plates for 30 min at 37°C. RNA was extracted using RNAzol B (Cinna/Biotecx Laboratories, Friendswood, TX). RNA (10 μg) was used in an oligo(dT)-primed cDNA synthesis using avian myeloblastosis virus reverse transcriptase. A 30-cycle PCR amplification was carried out with the primers F LARD Kpn, 5′-CGT AGG GTA CCC CAG GGC GGC ACT CGT AG, and R LARD Xba, 5′-GGC TCT AGA TCA CGG GCC GCG CTG CAG GC, which span exons 2–10. Products were cloned and sequenced, and a Southern blot was probed with ³²P-end-labeled primer F LARD Xba. A control PCR amplified a fragment from the glyceraldehyde-3-phosphate dehydrogenase cDNA.

Expression of LARD.

An expression vector, pCD33L, was constructed in which the CD33 leader sequence was cloned into pCDNA3 (Invitrogen). A clone expressing full-length LARD was constructed by cloning a PCR product generated using the primers F LARD Kpn and R LARD Not, 5′-GCC AAA CTG CGG CCG CGA AGT TGA G, into pCD33L. To make the Fas/LARD chimera, the Fas extracellular and transmembrane domains were cloned into the KpnI and EcoRI sites of CD33L using the primers F Fas Kpn, 5′-AAT GCG GTA CCT AGA TTA TCG TCC AAA AGT GTT AAT GCC C, and R FAS Eco, 5′-CAA GAA TTC TAC TTC CTT TCT CTT CAC CCA AAC AA. In a second step, the cytoplasmic domain of LARD was added by cloning a further PCR product generated with primers F LARD Eco, 5′-TGG GAA TTC ACC TAC ACA TAC CGC CAC TGC TGG, and R LARD Xba into the EcoRI and XbaI sites. Fas, TRAIL, and FasL expression vectors were constructed by cloning PCR products covering the entire coding regions of these genes into pCDNA3. Activity of FasL and TRAIL clones was checked by demonstrating that 293T transfectants could lead to apoptosis of Jurkat cells known to express both Fas and the receptor for TRAIL. CD8α, which was used in the cotransfection experiments, was cloned into the pREP-4 expression vector. 293T cells were transfected by calcium phosphate precipitation.

Tissue Distribution of LARD.

Northern blot analysis using a full-length LARD cDNA probe reveals a message of 3.5–4 kb (Fig. ​(Fig.2).2). In contrast to Fas and TNF-R1, LARD expression is restricted predominantly to lymphoid tissues, spleen, thymus, and PBL. To quantify LARD expression on PBL subsets, we developed an RNase protection assay using a 3′ portion of the LARD cDNA as a probe. LARD is expressed in roughly equal amounts in B cells and CD4⁺ or CD8⁺ T cells (Fig. ​(Fig.3).3). After PHA treatment of PBL, LARD mRNA expression is not increased. Interestingly, LARD is not expressed on macrophages, laboratory Epstein–Barr virus-transformed B cells, or the lymphocyte tumor cell lines (Jurkat, H9, HUT-78, C8166, Daudi, Raji, Namalwa, CIR, and HL-60), and very low level expression was seen on the T cell lines CEM and MOLT-4 and the colon cancer cell line HT-29.

Open in a separate window

Figure 2

Multiple-human-tissue Northern blot probed with LARD cDNA (Upper) and β-actin (Lower). Lanes: 1, spleen; 2, thymus; 3, prostate; 4, testis; 5, ovary; 6, small intestine; 7, colon; 8, PBL.

Open in a separate window

Figure 3

RNase protection analysis on lymphocyte subsets. (Upper) LARD expression; (Lower) control protection using a β-actin probe. Lanes: 1, control (no RNA); 2, unstimulated PBL; 3, phytohemagglutinin (PHA)-blasted PBL; 4, CD4⁺ cells; 5, CD8⁺ cells; 6, B cells; 7, macrophages.

Proapoptotic Activity of LARD.

Overexpression of Fas or TNF-R1 can lead to apoptosis in the absence of their ligands, probably through homomultimerization of the death domains (13). LARD was cotransfected with CD8α into 293T cells, and after 48 hr, the transfected cells showed the typical morphological features of apoptosis, and cell death was assessed by trypan blue staining. Expression of LARD resulted in an 18% increase in cell death above background (Fig. ​(Fig.44A). To test the effects of receptor crosslinking and also to check for function of the death domain, a fusion between the extracellular domain of Fas and the intracellular domain of LARD was made. Receptor crosslinking with an anti-Fas mAb leads to an increase in cell death from 20% to 35% for the Fas/LARD chimera compared with 50% for cells transfected with full-length Fas and exposed to anti-Fas antibody (Fig. ​(Fig.44B). The increased apoptosis seen with the full-length Fas construct probably results from 3-fold higher expression on transfected cells when compared with the Fas/LARD chimera (Fig. ​(Fig.44C).

Open in a separate window

Figure 4

Cell death was assessed by trypan blue staining and expressed as a percentage of the total. (A) Cells transfected with CD8 alone or cotransfected with CD8 and LARD. (B) effects of crosslinking with an anti-Fas mAb. F, Fas; L, LARD; F/L, Fas/LARD chimera. (C) Expression of Fas on transfected cells analyzed by a fluorescence-activated cell sorter. These data are representative of three separate experiments.

We have not yet been able to establish a stable cell line expressing LARD because of its propensity to cause spontaneous cell death. Neither the addition of TNF-α to LARD transfectants nor coculture with cells transfected with FasL or TRAIL leads to any increment in cell death (data not shown). This suggests that LARD is not the receptor for TNF-α, FasL, or TRAIL and has a novel, as-yet-unidentified ligand.

Alternative Pre-mRNA Splicing of LARD.

The first LARD clone we sequenced was an alternatively spliced variant containing a retained intron (LARD-11), which leads to the insertion of a premature termination codon. For this reason, we initiated a more systematic search for alternative splicing using reverse transcriptase–PCR to amplify LARD cDNA from resting lymphocyte subsets and PHA-blasted lymphocytes.

LARD is extensively alternatively spliced, producing at least 12 distinct variants (Fig. ​(Fig.5).5). Alignment of these clones with the cDNA encoding the full-length LARD protein, LARD-1, shows that all but four are generated by the skipping of one or more portions of LARD-1. Comparison of these clones with the genomic structure of Fas (14) suggests the LARD variants are generated by the skipping of one or more complete exons. This comparison allows a tentative scheme for the genomic structure of LARD to be proposed; this structure probably contains 10 exons, 6 for the leader and extracellular domain, 1 for the transmembrane domain, and 3 for the cytoplasmic domain (Fig. ​(Fig.55A).

Open in a separate window

Figure 5

(A) Schematic representation of the LARD protein drawn above the proposed genomic structure showing the leader peptide, four NGFR repeats, the transmembrane domain, and the death domain. (B) A proposed scheme for the genomic structure of LARD is represented above the prototype LARD-1 cDNA. Alternatively spliced clones LARD-2 through -11 are represented below LARD-1. Positions of the termination codons are represented by an asterisk above the clones. The CAG insertion in clone LARD-1b is indicated above LARD-1.

Clone LARD-2 and LARD-10 have 101- and 200-bp insertions leading to premature termination proximal to the sequence encoding the transmembrane domain. Both inserted sequences have matches to consensus splice sites at their 5′ and 3′ ends and, therefore, likely represent short retained introns. Finally, clone LARD-1b contains an extra three nucleotides, CAG, inserted at position 709, which probably reflects splicing to an alternative 3′ splice site three nucleotides proximal to the one used in LARD-1. The rest of the variants are generated by skipping one or more of exons 3–8. Most of the alternatively spliced clones have stop codons introduced before the transmembrane domain. Only three variants, LARD-7, -8, and -9, extend into the cytoplasmic domain; LARD-7 and -9 skip the transmembrane domain, whereas LARD-8, a low abundance clone, only lacks a single NGFR repeat. In summary, apart from LARD-1b and LARD-8 these variant cDNAs encode isoforms with either 2.5 or 3.5 NGFR repeats which are preceded by the normal leader sequence but have no transmembrane domain and would, therefore, be expected to be secreted in a soluble and probably functional form.

We thank U. Gerth for the CD8 plasmid, D. Simmons for the CD33 plasmid, S. Wilson for the HeLa cDNA library, and the Washington University–Merk EST Sequencing Project for the provision of plasmids. We also thank The Wellcome Trust, the Medical Research Council, the Arthritis and Rheumatism Council, and the Australian Government Commonwealth AIDS Research Grant for supporting the personnel and expenses of this research.