AND-34, a novel p130Cas-binding thymic stromal cell protein regulated by adhesion and inflammatory cytokines.

We have characterized a novel cDNA whose steady state mRNA levels rise in the thymus 2 to 6 h following the induction of CD4+CD8+ thymocyte apoptosis by in vivo cross-linking of CD3 epsilon. This cDNA, AND-34-1, contains an open reading frame (ORF) encoding a protein with an amino-terminal Src homology 2 (SH2) domain and a carboxyl-terminal domain homologous to GDP-exchange factors (GEFs). Northern analysis demonstrates widespread expression of the AND-34 gene. Anti-CD3 epsilon treatment induces up-regulation of the AND-34 mRNA levels in total thymic RNA but not in RNA from purified thymocytes, suggesting that this transcript is derived from a thymic stromal cell population. IL-1 and TNF increase AND-34 transcript levels in thymic cortical reticular, thymic nurse, and fibroblast cell lines. In the thymic cortical reticular cell line, IL-1 and TNF induce a protein of the predicted 93-kDa size reactive with anti-AND-34 peptide antisera. Fifteen minutes of serum stimulation of vanadate-pretreated AND-34-1-transfected NIH3T3 fibroblasts induces tyrosine phosphorylation of AND-34 as well as coprecipitating 95-, 125-, and 130-kDa proteins. One of these tyrosine phosphorylated proteins is identified as p130Cas (Crk-associated substrate), a signaling molecule previously known to bind to a GDP-exchange factor (C3G) and inducibly associate with the focal adhesion complex. Consistent with such an association, AND-34 tyrosine phosphorylation is induced following adherence of trypsinized fibroblasts to fibronectin or poly-L -lysine-coated surfaces.

Double-positive thymocytes resistant to antigen-MHC-induced negative selection lack active caspase.

Thymocytes bearing autoreactive TCR are eliminated from the organism by a process termed negative selection. The molecular basis of this deletion has been recently shown to be a consequence of TCR-triggered activation of a caspase by certain peptide-MHC ligands in the immature CD4+CD8+ double-positive (DP) thymocyte subpopulation. Of note, the numerically minor TCRhigh DP thymocyte subpopulation, unlike the major TCRlow DP subset, is resistant to negative selection. Despite exposure to cognate peptide, TCRhigh DP thymocytes mature into single-positive thymocytes and are exported into the periphery. Here we investigated the mechanism by which these thymocytes escape negative selection. Using a cytochemical assay in conjunction with a caspase-specific affinity ligand, we demonstrate that the resistance of the TCRhigh DP thymocytes to negative selection correlates with the disappearance of TCR-triggered caspase activity in these cells. Thus thymocytes which have presumably begun the positive selection process inactivate the thymic caspase pathway and are no longer susceptible to negative selection.

Pre-TCR signaling components trigger transcriptional activation of a rearranged TCR alpha gene locus and silencing of the pre-TCR alpha locus: implications for intrathymic differentiation.

A rearranged TCR alpha transgene remains transcriptionally inactive in rag-2-/- thymocytes but can be induced by CD3-mediated signals with concomitant maturation of double-negative (DN) thymocytes to the CD4+CD8+ double-positive (DP) stage. Reciprocally, the same signals silence pre-TCR alpha (pT alpha) expression. In normal C57BL/6 thymocytes, TCR alpha expression is not detected in DN thymocytes while, in contrast, TCR beta expression is initiated at the most immature c-kit+CD44+CD25- stage and continues throughout thymocyte development. pT alpha expression is first detected at the intermediate c-kit +/- CD44+CD25+ DN stage, increases during transition to the more mature c-kit-CD44-CD25+ stage and is lost at the DP stage. Thus, although TCR beta and pT alpha expression are independent, the pre-TCR complex mediates signals controlling the appearance of alpha beta TCR through selective regulation of TCR alpha and pT alpha genes.

Double-positive T cell receptor(high) thymocytes are resistant to peptide/major histocompatibility complex ligand-induced negative selection.

To investigate negative selection events during intrathymic ontogeny, we established T cell receptor (TCR)-transgenic mice [N15tg/RAG-2-/- (H-2b)] expressing a single TCR specific for vesicular stomatitis virus nuclear octapeptide N52-59 (VSV8) in the context of the major histocompatibility complex (MHC) class I molecule, K(b). Administration of VSV8 in vivo induced apoptosis in less than 4 h, deleting the majority of immature double-positive (DP) thymocytes by 24 h. In contrast, DP TCRhigh as well as single-positive (SP) thymocytes were refractory to this death process. Moreover, DP TCRhigh cells differentiated into SP thymocytes in vitro and in vivo, maturing into functional cytotoxic T lymphocytes upon intrathymic transfer to beta RAG 2-/- recipients. Hence, negative selection processes involving MHC-bound peptide ligands are operative only prior to the late DP thymocyte stage in this MHC class I-restricted TCR transgene system.

T-cell receptor ligation by peptide/MHC induces activation of a caspase in immature thymocytes: the molecular basis of negative selection.

T-cell receptors (TCRs) are created by a stochastic gene rearrangement process during thymocyte development, generating thymocytes bearing useful, as well as unwanted, specificities. Within the latter group, autoreactive thymocytes arise which are subsequently eliminated via a thymocyte-specific apoptotic mechanism, termed negative selection. The molecular basis of this deletion is unknown. Here, we show that TCR triggering by peptide/MHC ligands activates a caspase in double-positive (DP) CD4+ CD8+ thymocytes, resulting in their death. Inhibition of this enzymatic activity prevents antigen-induced death of DP thymocytes in fetal thymic organ culture (FTOC) from TCR transgenic mice as well as apoptosis induced by anti-CD3epsilon monoclonal antibody and corticosteroids in FTOC of normal C57BL/6 mice. Hence, a common caspase mediates immature thymocyte susceptibility to cell death.

Cross-linking of T-cell receptors on double-positive thymocytes induces a cytokine-mediated stromal activation process linked to cell death.

To investigate molecular events associated with the intrathymic process of negative selection, we established an in vivo system using an anti-CD3 epsilon monoclonal antibody to induce synchronous apoptosis in the thymus of AND T-cell receptor (TCR) transgenic RAG-2-/- mice in a non-selecting haplotype. This model eliminates endogenous negative selection as well as gene activation in the mature thymocyte compartment, offering an ideal source of tester (anti-CD3 epsilon-treated) and driver (untreated) thymus RNA for representational difference analysis (RDA). Fourteen mRNA sequences that are up-regulated in the thymuses of such mice 2-6 h after anti-CD3 epsilon treatment were identified. Surprisingly, the majority of these transcripts were derived from stromal cells rather than the TCR-cross-linked CD4+CD8+TCRlow thymocytes including the macrophage products IL-1, the chemokine Mig and the transcription factor LRG-21. IFN-gamma secretion from the CD4+CD8+TCRlow thymocytes regulates macrophage Mig production. Three other cytokines (IL-4, GM-CSF and TNF-alpha), known to activate a variety of stromal cells, are also induced in the same thymocyte population undergoing apoptosis. Expression of a TNF-alpha-inducible gene, B94, in stromal cells after TCR ligation further supports the notion of cross-talk between thymocytes and stroma. Thus, TCR-triggered immature thymocytes elaborate cytokines which may regulate the delivery of further signals from stromal cells required for apoptosis.

Delineation of a T-cell activation motif required for binding of protein tyrosine kinases containing tandem SH2 domains.

To define the T-cell receptor signal transduction motif, we have transfected human and murine T-cell lines with a chimeric receptor consisting of the extracellular and transmembrane domains of human CD8 alpha and the membrane-proximal portion of CD3 zeta containing at its C terminus either an 18-amino acid segment (NQLYNELNLGRREEYDVL) or alanine-scanning point mutant derivatives. Crosslinking of the extracellular domain of the chimera is sufficient to initiate Ca2+ flux, interleukin 2 production, and tyrosine phosphorylation of cellular proteins including the chimera. Subsequently, the chimera becomes associated with several tyrosine-phosphorylated proteins, among them the 70-kDa protein tyrosine kinase ZAP70. Mutational data identify the T-cell activation motif as Y(X)2L(X)7Y(X)2L and show that each of the four designated residues is necessary for the above activation events. Recombinant protein containing the two tandem SH2 domains derived from ZAP70 binds to a synthetic peptide corresponding to the above 18-amino acid motif but only when both tyrosines are phosphorylated; in contrast, little or no binding is observed to monophosphorylated or nonphosphorylated analogues. These results imply that after receptor crosslinking in T cells, and by inference also in B cells and mast cells, the motif is phosphorylated on both tyrosine residues, thereafter serving as a docking site for protein tyrosine kinases containing tandem SH2 domains.

Targeted disruption within the CD3 zeta/eta/phi/Oct-1 locus in mouse.

To elucidate the role of the CD3 eta subunit of the T cell receptor (TCR) in thymic development, a CD3 eta -/- mouse was generated by gene targeting. Insertion of a neomycin resistance gene into exon 9 of the CD3 zeta/eta/phi locus disrupted expression of CD3 eta and CD3 phi without affecting the expression of CD3 zeta. Little difference was observed between wild type and CD3 eta -/- mice with regard to cellularity or subset composition in thymus and peripheral lymphoid organs. Furthermore, neither alloproliferative responses nor cytotoxic T lymphocyte generation and effector function was affected by the mutation. The effect of the CD3 eta -/- mutation on thymic selection was examined by crossing the CD3 eta knockout animals with anti-HY TCR transgenic animals: the absence of the CD3 eta subunit altered neither positive nor negative selection. Thus, CD3 eta is not required for thymic selection. Of note, the birth rate of the CD3 eta -/- animals was significantly lower than that of wild type or heterozygous animals (P = 0.041-0.002). This unexpected result is probably the consequence of an alteration in mRNA expression of the Oct-1 nuclear transcription factor in CD3 eta -/- animals. The CD3 zeta/eta/phi locus partially overlaps the gene encoding Oct-1 whose transcription is dysregulated by the CD3 eta -/- mutation. Our results clearly underscore the value of characterizing all products of a genetic locus disrupted by gene targeting.

CD3 zeta/eta/theta locus is colinear with and transcribed antisense to the gene encoding the transcription factor Oct-1.

CD3 zeta and eta are signal-transducing components of the TCR and are derived from alternative splicing of transcripts from a single genetic locus that also encodes CD30 theta. We have isolated two murine cDNA clones that appear to result from antisense transcription through CD3 theta-specific exon 10 and CD3 eta-specific exon 9. The sequence of these clones shows no open reading frame. Northern analysis with single stranded probes confirms the existence of a ubiquitously expressed > 12-kb polyadenylated mRNA antisense to CD3 eta. A "genomic walk," which extended 32 kb distal to murine CD3 eta exon 9, provided genomic DNA containing a more 5' portion of the antisense transcript. This probe identified two murine thymic cDNA with 91% sequence homology to the human transcription factor Oct-1. Five exons of murine Oct-1 map in an antisense orientation to the CD3 zeta/eta/theta locus on the cloned genomic sequences. The murine Oct-1 cDNA and exon 9 of CD3 eta hybridize to the same > 12-kb mRNA. Similarly, human Oct-1 and previously characterized human genomic sequences homologous to murine CD3 eta exon 9 each hybridize to the same > 15-kb human mRNA. Thus, the CD3 zeta/eta/theta and Oct-1 gene loci are partially overlapping and transcribed in opposite directions. The potential functional implications of these findings are discussed.

Overexpression of CD3 eta during thymic development does not alter the negative selection process.

To elucidate the role of CD3 eta in thymic development and to determine whether CD3 eta is involved in the negative selection process, CD3 eta was overexpressed > 100 fold in transgenic (tg) mice using a Thy-1 promoter and regulatory elements. CD3 eta was readily observed in the majority of cortical thymocytes and in a fraction of medullary thymocytes in tg mice by immunohistochemical staining with an anti-CD3 eta-specific mAb. In contrast, endogenous CD3 eta levels were too low to detect in normal littermates. Flow cytometric analysis demonstrated an increased level of TCR on thymocytes with intermediate TCR density in tg animals and parallel biochemical studies showed a marked increased in TCR-associated CD3 zeta-eta heterodimers and CD3 eta-eta homodimers relative to controls. Despite this change in surface TCR phenotype, there was no significant alteration in the total numbers or proportion of CD4+CD8+ double-positive or CD4+CD8- or CD4-CD8+ single-positive thymocytes or peripheral T cells. Percentages of SP V beta 5, V beta 6, and V beta 8 thymocytes in tg animals were unaltered compared to normal littermates when backcrossed either to C57BL/6 (H-2b) or DBA/2 (H-2d) backgrounds. Furthermore, induction of DNA fragmentation with anti-CD3 epsilon mAb treatment in vivo was not significantly different for tg and normal littermates. Collectively, these data imply that CD3 eta is not a limiting component of the negative selection process.

Differential regulation of T-cell receptor processing and surface expression affected by CD3 theta, an alternatively spliced product of the CD3 zeta/eta gene locus.

The T-cell receptor (TCR) is a multisubunit complex consisting of the clonotypic Ti alpha and beta (or Ti gamma and delta) subunits and the invariant CD3 gamma, CD3 delta, CD3 epsilon, CD3 zeta, and CD3 eta subunits. Herein, we describe an additional product from the CD3 zeta/eta gene locus which we have termed CD3 theta. The cDNA derives from the first seven exons common to CD3 zeta and CD3 eta, 94 base pairs (bp) of the CD3 eta-specific exon 9 and an additional exon 10 encoding the carboxyl-terminal 15 amino acids and the 3'-untranslated region. The expression of CD3 theta is equivalent to that of CD3 eta in tissue distribution and level of expression as judged by RNase protection analysis. Despite the identity of the amino-terminal 121 amino acids of CD3 zeta, CD3 eta, and CD3 theta and an additional 31 amino acids shared between CD3 eta and CD3 theta, transfection of CD3 theta into the CD3 zeta- eta- T-cell hybridoma, MA5.8, failed to restore detectable surface TCR expression in contrast to transfection with CD3 zeta or CD3 eta. Analysis of the CD3 theta protein in transfectants indicated that CD3 theta is associated with the TCR intracellularly. However, unlike with CD3 zeta, Ti alpha-beta chains remain endoglycosidase H sensitive, suggesting a role for the unique COOH-terminal segment of CD3 theta in mediating TCR retention and/or degradation in a pre-Golgi compartment.

Human immunodeficiency virus gp120 binding C'C" ridge of CD4 domain 1 is also involved in interaction with class II major histocompatibility complex molecules.

Using site-directed mutagenesis informed by high-resolution CD4 structural data, we have investigated the role of residues of the C'C'' ridge region of human CD4 on class II major histocompatibility complex (MHC) binding. This C'C'' ridge is homologous to the CDR2 loop of an immunoglobulin variable domain and is known to contain the binding site for human immunodeficiency virus (HIV) coat glycoprotein gp120. Here we report that this region is also involved in interaction with class II MHC. Exposed positively charged residues Lys-35, Lys-46, and Arg-59 and the exposed hydrophobic residue Phe-43 contribute significantly to class II MHC binding. Moreover, mutations in the buried residues Trp-62 and Ser-49, which support the top and bottom of the C'C'' ridge, respectively, disrupt class II MHC interaction. The HIV binding region appears to involve a restricted area of the larger class II MHC binding site on CD4. Strategies of drug design aimed at interrupting CD4-HIV interaction will need to consider the extensive overlap between class II MHC and HIV gp120 binding surfaces in this region of CD4.

Negative selection of thymocytes. A novel polymerase chain reaction-based molecular analysis detects requirements for macromolecular synthesis.

Self-tolerance is mainly established through clonal deletion of autoreactive T cells during thymic differentiation. The mechanisms by which deletion is achieved are poorly understood. Here we use a specific polymerase chain reaction-based system to characterize DNA fragmentation and show that after in vivo treatment of neonatal mice with staphylococcus enterotoxin B, selective apoptosis of V beta 8+ thymocytes occurs. This process precedes detectable deletion of V beta 8+ cells as determined by phenotypic analysis. Moreover, in vivo administration of cycloheximide and, to a lesser extent, actinomycin D, inhibits apoptosis of staphylococcus enterotoxin B specific thymocytes. Thus, macromolecular synthesis is a requirement for negative selection.

The human immunodeficiency virus gp120 binding site on CD4: delineation by quantitative equilibrium and kinetic binding studies of mutants in conjunction with a high-resolution CD4 atomic structure.

The first immunoglobulin V-like domain of CD4 contains the binding site for human immunodeficiency virus gp120. Guided by the atomic structure of a two-domain CD4 fragment, we have examined gp120 interaction with informative CD4 mutants, both by equilibrium and kinetic analysis. The binding site on CD4 appears to be a surface region of about 900 A2 on the C" edge of the domain. It contains an exposed hydrophobic residue, Phe43, on the C" strand and four positively charged residues, Lys29, Lys35, Lys46, and Arg59, on the C, C', C", and D strands, respectively. Replacement of Phe43 with Ala or Ile reduces affinity for gp120 by more than 500-fold; Tyr, Trp, and Leu substitutions have smaller effects. The four positively charged side chains each make significant contributions (7-50-fold). This CD4 site may dock into a conserved hydrophobic pocket bordered by several negatively charged residues in gp120. Class II major histocompatibility complex binding includes the same region on CD4; this overlap needs to be considered in the design of inhibitors of the CD4-gp120 interaction.

Human genomic sequences corresponding to murine CD3 eta-related transcripts: lack of conservation or expression of homologous human products.

We have cloned and sequenced human genomic DNA homologous to exons 9 and 10 of the CD3 zeta/eta/theta locus. Although there are open reading frames within the human sequences corresponding to the translated portions of murine exons 9 and 10, we find no evidence of conservation of the encoded polypeptide product. Furthermore, using oligonucleotides derived from these homologous sequences, we are unable to detect human CD3 eta- or CD3 theta-like transcripts by polymerase chain reaction amplification of reverse-transcribed RNA from a variety of human lymphoid tissues. Despite the absence of evidence for conservation of human CD3 eta and CD3 theta, there is a surprising degree of similarity between human and murine nucleotide sequences, not only for exons 9 and 10 (78% and 70%, respectively), but also for the 9/10 intron (71%). A possible mechanism for this conservation is discussed.

Phosphorylation of multiple CD3 zeta tyrosine residues leads to formation of pp21 in vitro and in vivo. Structural changes upon T cell receptor stimulation.

T lymphocyte activation resulting from antigen recognition involves a protein tyrosine kinase pathway which triggers phosphorylation of several cellular substrates including the CD3 zeta subunit of the T cell receptor (TCR) to form pp21. The homologous TCR-associated protein, CD3 eta, is an alternatively spliced product of the same gene locus as CD3 zeta. CD3 eta lacks one of six cytoplasmic tyrosine residues (Tyr-132) found in CD3 zeta and is itself not phosphorylated. Site-directed mutagenesis in conjunction with in vitro and in vivo phosphorylation studies herein demonstrates that Tyr-132 is required for the formation of pp21. Moreover, the differential phosphorylation of CD3 zeta versus CD3 eta is not due to a selective association of the known TCR-associated protein tyrosine kinase, p59fyn; p59fyn but not p56lck or p62yes is associated with each of the three TCR isoforms containing CD3 zeta 2, or CD3 eta 2, or CD3 zeta-eta. This association occurs through components of the TCR complex distinct from CD3 zeta or CD3 eta. In addition, we show that pp21 formation is not only dependent on Tyr-132 but results from concomitant phosphorylation of other CD3 zeta residues including Tyr-121. Mutation of Tyr-90, -121, or -132 does not alter primary signal transduction as shown by the ability of individual CD3 zeta Tyr----Phe mutants to produce interleukin-2 upon TCR stimulation. Thus, the substantial structural changes in CD3 zeta upon TCR stimulation as reflected by alteration in its mobility in sodium dodecyl sulfate-polyacrylamide gel electrophoresis may affect subsequent events such as receptor desensitization, receptor movement, and/or protein associations.

T cell receptor complexes containing Fc epsilon RI gamma homodimers in lieu of CD3 zeta and CD3 eta components: a novel isoform expressed on large granular lymphocytes.

CD3 zeta and CD3 eta form disulfide-linked homo- or heterodimers important in targeting partially assembled Ti alpha-beta/CD3 gamma delta epsilon T cell receptor (TCR) complexes to the cell surface and transducing stimulatory signals after antigen recognition. Here we identify a new TCR isoform expressed on splenic CD2+, CD3/Ti alpha-beta+, CD4-, CD8-, CD16+, NK1.1+ mouse large granular lymphocytes (LGL), which are devoid of CD3 zeta and CD3 eta proteins. The TCRs of this subset contain homodimers of the gamma subunit of the high affinity receptor for IgE (Fc epsilon RI gamma) in lieu of CD3 zeta and/or CD3 eta proteins. The LGL display natural killer-like activity and are cytotoxic for B cell hybridomas producing anti-CD3 epsilon and anti-CD16 monoclonal antibodies, demonstrating the signaling capacity of both TCR and CD16 in this cell type. These findings provide evidence for an additional level of complexity of TCR signal transduction isoforms in naturally occurring T cell subsets.

T-cell-receptor isoforms.

Early work on T-cell hybridomas lacking the T-cell-receptor (TCR) sub-unit CD3 eta had suggested a correlation between the presence of CD3 zeta-eta heterodimers and signalling leading to phosphatidyl-inositol (PI) turnover as well as activation-induced cell death. The cloning of CD3 eta has now allowed thorough and direct analysis of the signal transduction properties of CD3 zeta-zeta-, CD3 zeta-eta- and CD3 eta-eta-containing TCRs. We have found that all forms of the TCR are capable of transducing signals leading to PI turnover, Ca2+ mobilization, IL-2 production and cell-cycle arrest. CD3 zeta and CD3 eta utilize the same promoter which yields coordinate expression of both products, so that restricted CD3 eta expression in a sub-population of thymocytes is unlikely. Immunohistochemical methods employing an anti-CD3 eta-specific monoclonal antibody (MAb) show no detectable staining of thymic sections from adult mice, implying at best a low level of constitutive CD3 eta expression. In contrast, CD3 eta expression is readily detected in the majority of cortical thymocytes of CD3 eta transgenic mice using a Thy-1 promoter construct. However, over-expression of CD3 eta in mice transgenic for this polypeptide does not result in increased negative selection in vivo, consistent with the in vitro findings that induction of cell death is not strictly dependent on CD3 eta. Despite earlier reports of the detection of human CD3 eta protein, we find no CD3 eta message in human thymus or T cells. Cloning of the human CD zeta-eta genomic locus has demonstrated approximately 70% homology between the mouse and human genomic sequence, corresponding to the mouse CD3 eta-specific exon. However, translation of the DNA sequence does not result in a homologous amino acid sequence. Thus, there does not appear to be a CD3 eta protein in humans.

Cellular myosin heavy chain in human leukocytes: isolation of 5' cDNA clones, characterization of the protein, chromosomal localization, and upregulation during myeloid differentiation.

We have isolated 5' cDNA clones encoding a member of the cellular myosin heavy chain gene family from human leukocytes. The predicted amino acid sequence shows 93% identity to a chicken cellular myosin heavy chain, 76% to chicken smooth muscle, and 40% to human sarcomeric myosin heavy chain. The mRNA is expressed as a 7.4- to 7.9-kb doublet in many nonmuscle cells, and is upregulated in myeloid cell lines on induction from a proliferating to a differentiated state. Antisera raised against a peptide made from the predicted amino acid sequence specifically reacts with a 224-Kd polypeptide in leukocyte cell lines, and the protein is also upregulated during the induction of monocytic and granulocytic differentiation in these cells. The gene for this cellular myosin heavy chain maps to chromosome 22, bands q12.3-q13.1, demonstrating that it is not located in the previously described sarcomeric gene clusters on chromosomes 14 and 17. This cellular myosin heavy chain may be a major contractile protein responsible for movement in myeloid cell lines because no mRNA for sarcomeric myosin heavy chain is detected in these cells.

CD3 eta and CD3 zeta are alternatively spliced products of a common genetic locus and are transcriptionally and/or post-transcriptionally regulated during T-cell development.

The CD3 eta subunit of the T-cell receptor is thought to subserve an important role in signal transduction and possibly T-cell development. Herein we characterize the organization of the mouse CD3 eta gene and show that it is part of one gene locus that also encodes CD3 zeta on chromosome 1. The NH2-terminal sequence of CD3 zeta and CD3 eta, which share the same leader peptide and are identical through amino acid 122 of each mature protein, is encoded by exons 1-7. However, exons 8 and 9 are differentially spliced to give rise to CD3 zeta and CD3 eta: exons 1-8 encode CD3 zeta and exons 1-7 plus 9 encode CD3 eta. RNase protection analysis with RNA from a variety of fetal, neonatal, and adult cell types indicates that expression of both gene products is T-lineage-restricted. Importantly, expression of CD3 zeta and CD3 eta mRNA appears before or on day 16 of fetal gestation. Expression is apparently coordinate since no cell types tested express CD3 zeta or CD3 eta alone. The steady-state level of CD3 zeta mRNA is greater than or equal to 40-60 times that of CD3 eta mRNA. In immature CD4+CD8+CD3low double-positive thymocytes and CD4+CD8-CD3high or CD4-CD8+CD3high single-positive thymocytes, the respective steady-state CD3 zeta and CD3 eta mRNA levels are equivalent, whereas the amount of receptor-associated CD3 zeta and CD3 eta proteins in double-positive thymocytes is approximately 10 times less than in single-positive thymocytes. Nevertheless, the CD3 zeta/CD3 eta protein ratio remains constant in all populations (40-60:1). Furthermore, discordance between mRNA and protein levels for CD3 zeta and CD3 eta is also observed in splenic T cells. Thus, posttranscriptional and/or transcriptional regulatory mechanisms control CD3 zeta and CD3 eta expression during T-cell development.