The cytoplasmic amino-terminus of the Latent Membrane Protein-1 of Epstein-Barr Virus: relationship between transmembrane orientation and effector functions of the carboxy-terminus and transmembrane domain


The Latent Membrane Protein 1 (LMP-1) protein of Epstein-Barr virus (EBV) is localized in the plasma membrane of the infected cell. LMP-1 possesses a hydrophobic membrane spanning domain, and charged, intracellular amino- and carboxy-termini. Two models have been proposed for the contribution of the amino-terminus to LMP-1's function: (i) as an effector domain, interacting with cellular proteins, or (ii) as a structural domain dictating the correct orientation of transmembrane domains and thereby positioning LMP-1's critical effector domains (i.e. the carboxy-terminus). However, no studies to date have addressed directly the structural contributions of LMP-1's cytoplasmic amino-terminus to function. This study was designed to determine if LMP-1's cytoplasmic amino-terminus (N-terminus) encodes information required solely for maintenance of proper topological orientation. We have constructed LMP-1 chimeras in which the cytoplasmic N-terminus of LMP-1 is replaced with an unrelated domain of similar size and charge, but of different primary sequence. Retention of the charged amino-terminal (N-terminal) cytoplasmic domain and first predicted transmembrane domain was required for correct transmembrane topology. The absolute primary sequence of the cytoplasmic N-terminus was not critical for LMP-1's cytoskeletal association, turnover, plasma membrane patching, oligomerization, Tumor Necrosis Factor Receptor-associated factor (TRAF) binding, NF-κB activation, rodent cell transformation and cytostatic activity. Furthermore, our results point to the hydrophobic transmembrane domain, independent of the cytoplasmic domains, as the primary LMP-1 domain mediating oligomerization, patching and cytoskeletal association. The cytoplasmic amino-terminus provides the structural information whereby proper transmembrane orientation is achieved.


EBV, the etiological agent of infectious mononucleosis, is a human herpesvirus involved in the development of several human malignancies (Miller, 1985). EBV's contribution to lymphoid malignancies in vivo is reflected in its ability to immortalize primary human B-cell in vitro. EBV is maintained in the nucleus of the immortalized B-cell as a replicating plasmid from which a subset of latent viral genes are expressed. LMP-1 is of particular interest because it is an EBV-encoded transforming gene (Baichwal and Sugden, 1988; Wang et al., 1985) essential for B-cell immortalization (Kaye et al., 1993) and is expressed in tumor cells of many EBV-associated malignancies (Fahreus et al., 1988; Gratama et al., 1991; Pallesen et al., 1991; Thomas et al., 1990; Young et al., 1989).

LMP-1 is a polytopic plasma membrane protein likely to function as a constitutively-active cell surface receptor (Fennewald et al., 1984; Gires et al., 1997; Mann et al., 1985). Like cell surface receptors (Gross et al., 1983; Haigler et al., 1978; Schlessinger et al., 1978; Yarden and Ullrich, 1988), LMP-1 and its functional deletion mutants localize at the cell surface in aggregated patches (Liebowitz et al., 1986; Mann et al., 1985; Martin and Sugden, 1991b), associate with the cytoskeleton (Liebowitz et al., 1987; Mann and Thorley-Lawson, 1987; Martin and Sugden, 1991b), and turn over rapidly in the cell (Baichwal and Sugden, 1987; Mann and Thorley-Lawson, 1987; Martin and Sugden, 1991b). LMP-1's tertiary structure, biochemical properties, interaction with cellular signaling proteins (Mosialos et al., 1995), and self-oligomerization (Eliopoulos and Rickinson, 1998; Eliopoulos and Young, 1998; Floettmann and Rowe, 1997; Gires et al., 1997; Hatzivassiliou et al., 1998) are consistent with LMP-1 functioning as a constitutively active cell surface receptor. LMP-1 regulates cellular signaling pathways coupled to activation of NF-κB, AP-1, and STAT transcription factors (Eliopoulos and Young, 1998; Floettmann and Rowe, 1997; Fries et al., 1996; Hammerskjold and Simurda, 1992; Hatzivassiliou et al., 1998; Huen et al., 1995; Kieser et al., 1997; Laherty et al., 1992; Mitchell and Sugden, 1995; Paine et al., 1995; Sylla et al., 1998) via interaction of its carboxy-terminal (C-terminal) activation domains 1 and 2 (CTAR-1 and -2) with intracellular TRAF and TRADD signaling molecules (Brodeur et al., 1997; Devergene et al., 1996, 1998; Eliopoulos et al., 1999; Izumi et al., 1997, 1999; Izumi and Kieff, 1997; Kaye et al., 1996; Kieser et al., 1999; Miller et al., 1998; Mosialos et al., 1995; Sandberg et al., 1997).

LMP-1's predicted tertiary structure includes a hydrophobic membrane-associated domain consisting of six probable transmembrane helices, a positively charged N-terminus of 20 residues and an acidic C-terminus of 190 residues, both of which are intracellular (Bankier et al., 1983; Liebowitz et al., 1986). All three domains are required for most of LMP-1 biological activities and biochemical properties (Huen et al., 1995; Izumi et al., 1994; Kaye et al., 1995; Liebowitz et al., 1992; Mitchell and Sugden, 1995; Moorthy and Thorley-Lawson, 1993; Wang et al., 1988b), with some exceptions (Baichwal and Sugden, 1989; Hammerschmidt et al., 1989; Martin and Sugden, 1991b). The N-terminal 25 residues of LMP-1 are essential for growth transformation, NF-κB activation, rapid turnover, cytoskeletal association, and membrane patching (Baichwal and Sugden, 1989; Hammerschmidt et al., 1989; Huen et al., 1995; Liebowitz et al., 1992; Martin and Sugden, 1991b; Mitchell and Sugden, 1995; Moorthy and Thorley-Lawson, 1993; Wang et al., 1988b). Without this cytoplasmic domain, or smaller proline/arginine-rich subdomains of it, LMP-1 is greatly impaired in its ability to contribute to B-cell immortalization by EBV (Izumi et al., 1994). Mutational analyses to date have not discriminated between potential structural and/or effector functions of the cytoplasmic N-terminus. Because it precedes the first transmembrane domain and has a net positive charge, the N-terminus may possess topological information that acts in concert with the likely signal-anchoring function of the first transmembrane domain to determine LMP-1's transmembrane topology (Hartmann et al., 1989; Lipp and Dobberstein, 1986; Parks and Lamb, 1991). These 20 amino acids could potentially encode effector functions as well. Results of deletion analyses are consistent with both models (membrane anchoring or effector domain) and the two models needn't be mutually exclusive. However, potential structural, charge-based functions cannot be dissociated from possible effector functions by deletion analyses because charged residues likely to play a role in topology are distributed throughout the 20 amino acid N-terminus.

To determine experimentally if LMP-1's cytoplasmic N-terminus encodes information required solely for maintenance of proper topological transmembrane orientation, we have constructed chimeras in which LMP-1's cytoplasmic N-terminus is replaced with unrelated domains of similar size and positive net charge, but of unrelated primary sequence. Results of biochemical and functional studies with these chimeras are the first to demonstrate definitively a structural role for the cytoplasmic N-terminus and are consistent with a model in which the cytoplasmic N-terminus, in the context of an intact first transmembrane domain, functions primarily as a topological domain and thereby contributes to LMP-1's known effector functions.


AGSPR/LMP-1 chimeras

LMP-1 substitution mutants were generated in which the cytoplasmic N-terminus was replaced with heterologous domains possessing positive net charges ranging from +0.5 to +2.0 (the net charge of LMP-1's cytoplasmic N-terminus is +2.5). The four LMP-1 substitution mutants encode a portion of the cytoplasmic N-terminus of the asialoglycoprotein receptor H1 (ASGPR) in place of LMP-1's cytoplasmic N-terminus (Figure 1a,b). The ASGPR is a single-spanning membrane protein with a cytoplasmic N-terminal domain of 40 residues and a large, glycosylated extracellular C-terminus of 231 residues (Figure 1a) (Spiess et al., 1985). ASGPR is a C-type lectin of hepatocytes that recognizes desialylated glycoproteins for endocytosis and lysosomal degradation. Like LMP-1, ASGPR has no cleavable signal peptide, but instead is targeted to the endoplasmic reticulum by an internal hydrophobic segment (Spiess and Lodish, 1986). The role of the ASGPR N-terminus in H1 receptor biogenesis has been studied extensively and the contribution of its charged residues to proper transmembrane orientation (Ncyt/Cexo) is well documented (Beltzer et al., 1991). Like LMP-1's N-terminus, the N-terminus of the ASGPR has a positive net charge (+1.5), but lacks LMP-1's proline/arginine rich subdomain (Figure 1b; see Materials and methods for sequence). This sequence, or subportions of it, should provide the necessary topological information for proper transmembrane orientation of LMP-1. Functional complementation of LMP-1's cytoplasmic N-terminus by this heterologous domain would strongly support a model in which LMP-1's N-terminus functions primarily as a topological domain.

Figure 1

(a) Transmembrane topology of LMP-1 (predicted), ASGPR H1 (proven topology), and ASGPR/LMP-1 chimeras (predicted). The N-termini of the shown proteins are intracellular. (b) Comparison of N-terminal region of LMP-1 and chimeras used in this study. Only the cytoplasmic N-terminus and first transmembrane domain are shown. The net charge of the cytoplasmic N-terminus is indicated for each protein. Also shown is the charge difference Δ(C-N) (Hartmann et al., 1989) between the sequences on the N-terminal and C-terminal sides of the transmembrane domain. (c) Primary sequence comparison between LMP-1 and mycASGPR/LMP-1. Only the sequence of the cytoplasmic N-terminus and first transmembrane domain of each protein is shown. Sequences from the ASGPR H1 receptor and myc tag are delineated, as is the proline-arginine rich region in LMP-1's N-terminus. The net charge of the cytoplasmic domain is indicated below each schematic. (d) Expression of LMP-1 chimeras. Western analysis of 293 cells transfected with CMV-promoter driven expression vectors using anti-LMP-1 antisera raised against LMP-1's C-terminus. Top blot, extracts of cells transfected with the following plasmids: (1) pCMV-LMP-1; (2) pCMV-mycASGPR/LMP-1; (3) pCMV-NΔ25; (4) pCMV-ASGPR/LMP-1.4; (5) pCMV-ASGPR/LMP-1.17; (6) pCMV-ASGPR/LMP-1.20; Bottom blot, extracts from cells transfected with pCMV-LMP-1 or pCMV-mycASGPR/LMP-1. WCL, whole cell lysate of transfected 293 cells; myc IP, anti-myc immunoprecipitates of extracts from transfected 293 cells. Arrows point to the wild type LMP-1 protein. The asterisk marks the position of the major cross-reactive LMP-1 protein comigrating with the LMP-1 deletion mutant NΔ43. B958, B958 cell extract included as a marker for the migration of LMP-1

The four ASGPR/LMP-1 chimeras and an N-terminal LMP-1 deletion mutant were expressed in 293 cells. NΔ25 is a nonfunctional LMP-1 deletion mutant lacking the first 25 codons from the N-terminus (Baichwal and Sugden, 1989; Hammerschmidt et al., 1989; Martin and Sugden, 1991b). In addition to lacking LMP-1's cytoplasmic N-terminal 20 residues, ASGPR/LMP-1 chimeras (.20, .17, .4) lack residues 21–25 (LSSSL) which are predicted to extend into LMP-1's first transmembrane domain (Materials and methods, and Figure 1). The mycASGPR/LMP-1 chimera differs from the other ASGPR/LMP-1 chimeras in that it encodes a myc tag at the N-terminus and only 10 residues from the ASGPR (ASGPR residues 16–25) (Figure 1c). Importantly, the N-terminal LSSSL sequence in the LMP-1's first transmembrane domain is retained in mycASGPR/LMP-1 so that only residues 1–20 of LMP-1 are replaced by heterologous sequences.

LMP-1-immunoreactive species migrating with the predicted molecular weight of the ASGPR/LMP-1 chimeras were detected, as well as faster migrating species. The predominant smaller molecular weight species (marked by asterisk in Figure 1d) comigrated with the LMP-1 deletion mutant NΔ43 which initiates at a methionine codon immediately proceeding the first transmembrane domain (Baichwal and Sugden, 1988; Hammerschmidt et al., 1989; Martin and Sugden, 1991b, and not shown) (Figure 1d). This species was frequently present at levels approaching that of the full length LMP-1 species (see lanes 3 (NΔ25), 5 (ASGPR/LMP-1.17), and 6 (ASGPR/LMP-1.20), Figure 1d). Truncation of LMP-1's N-terminal transmembrane domain (as in ASGPR/LMP-1 .17, .20 and .4, and NΔ25) may result in proteolytic cleavage at a cryptic signal peptidase cleavage site following the first transmembrane domain, giving rise to the faster migrating LMP-1 species (Schmid and Spiess, 1988; Spiess and Lodish, 1986) (Figure 1d). Consistent with its predicted size and the presence of the N-terminal myc tag, the predominant LMP-1-immunoreactive protein product detected following expression of mycASGPR/LMP-1 comigrated with full length LMP-1 and was specifically immunoprecipitated by anti-myc antibodies (Figure 1d). Rarely did we observe the truncated product comigrating with NΔ43 following expression of mycASGPR/LMP-1.

Transmembrane orientation of chimeras

The contribution of the cytoplasmic N-terminus to LMP-1's biogenesis was examined by comparing the transmembrane orientation of the LMP-1 variants to that of wild type LMP-1 following transfection of expression vectors into 293 cells. The transmembrane orientation of chimeras and wild type LMP-1 was determined by incubating intact transfected cells with the appropriate protease. Cleavage products were visualized by Western analysis using antibodies that recognize LMP-1's intracellular C-terminus. LMP-1 variants exhibiting wild type LMP-1 transmembrane orientation will yield C-terminal cleavage products indistinguishable from LMP-1's. Predicted cleavage sites for three proteases are shown in Figure 2a. The mycASGPR/LMP-1 chimera was immunoprecipitated from trypsinized intact cells with anti-myc antibodies, indicating that the N-terminus was protected from cleavage with exogenous trypsin and therefore was intracellular (the lack of an N-terminal epitope tag in ASGPR/LMP-1 .4, .17, .20 and NΔ25 precluded their being tested in this fashion). Likewise, the cytoplasmic C-terminus of mycASGPR/LMP-1 was protected from tryptic cleavage and thus was intracellular (Figure 2b). Identical results were obtained with N-terminally tagged wild type LMP-1 under the same conditions (not shown). The intracellular localization of the N- and C-termini of mycASGPR/LMP-1 was confirmed by comparison of anti-myc and anti-LMP-1 immunoreactivity in permeabilized and unpermeabilized cells using single cell immunofluorescent analysis (Table 1).

Figure 2

Transmembrane orientation of LMP-1 and LMP-1 chimeras in 293 cells. (a) Schematic of LMP-1 showing location of predicted protease cleavage sites. No trypsin cleavage sites in LMP-1 are exposed to the outside of the cell. The only predicted chymotrypsin (Liebowitz et al., 1986; Martin and Sugden, 1991a) and pronase cleavage sites accessible to cleavage in intact cells are indicated by the arrows. Substituted N-terminal domains of all chimeras possess between 5–6 predicted trypsin cleavage sites, 1–2 predicted chymotrypsin cleavage sites (except mycASGPR/LMP-1 which has none), and 5–8 predicted pronase cleavage sites. (b–d) Sensitivity of LMP-1 chimeras to trypsin (b), chymotrypsin (c), and pronase (d) cleavage. Intact 293 cells (electroporated with CMV-based expression vectors) were treated 24 h post-transfection with proteases as described in Materials and methods. Extracts were resolved by SDS–PAGE and analysed by Western blot using anti-LMP-1 antisera. The LMP-1 protein expressed from the introduced vector is indicated at the top of each blot. mycIP, anti-myc immunoprecipitation of extracts from trypsinized (trypsin; 1X or 0.3X), or untrypsinized (−) cells (b only); B958, extract of the EBV-positive B958 cell line included as a marker for the migration of LMP-1 proteins; the specific protease used is indicated to the left of each blot: (−), incubation with PBS (no protease); (1X), undiluted protease, (0.3X), 1 : 3 dilution of protease; the asterisk indicates the position of the major proteolytic cleavage products for each protein

Table 1 Table 1

The only predicted chymotrypsin and pronase cleavage sites in LMP-1 accessible to cleavage in intact cells are located in the loop between the first and second putative transmembrane domains (Figure 2a) (Liebowitz et al., 1986; Martin and Sugden, 1991a). The transmembrane topology of mycASGPR/LMP-1 was indistinguishable from that of LMP-1 as shown by identical chymotrypsin and pronase digestion patterns (Figure 2c,d). The ASGPR/LMP-1 chimeras (.4, .17, .20) and the NΔ25 deletion mutant were improperly inserted into the plasma membrane as evidenced by the generation of C-terminal cleavage products migrating faster than those from LMP-1 (Figure 2c,d). Presumably these products arose by cleavage of exposed sites which, in LMP-1, are localized within the cell and thus protected from cleavage. Faster migrating cleavage products are represented in LMP-1 control extracts at low levels. Generation of such cleavage products from LMP-1 likely results from a combination of intracellular proteolysis of full length protein (Baichwal and Sugden, 1987) due to its improper folding or insertion into the ER membrane (Bonifacino and Lippincott-Schwartz, 1991; Bonifacino and Weissman, 1998), or as a result of cleavage of improperly inserted and transported LMP-1 by applied protease (Hedge and Lingappa, 1997, 1999). Interestingly, despite their aberrant topology, the N- and C-termini of these mutants were intracellular as shown by immunological analyses of trypsinized intact cells (not shown). Thus, properly localized termini in such a multiple-membrane spanning protein may not necessarily be sufficient for specification of wild type transmembrane orientation of membrane spanning domains. The ASGPR/LMP-1 chimeras differ primarily from mycASGPR/LMP-1 in that they lack the LSSSL sequence at the start of the first transmembrane domain (see Figure 1 and Materials and methods). Interestingly, the ASGPR/LMP-1 chimeras (.4, .17, .20) behaved very much like the nonfunctional NΔ25 deletion mutant in a variety of functional and biochemical assays (not shown and see below) and their altered topology is considered in the interpretation of subsequent functional assays. These results demonstrate a critical role for the cytoplasmic N-terminus and first transmembrane domain in determining LMP-1's transmembrane topology, and this domain can be effectively replaced with a similarly charged, unrelated domain. We focused primarily on the mycASGPR/LMP-1 chimera for the majority of the following studies because, like LMP-1, mycASGPR/LMP-1 is expressed primarily in a full length, unproteolyzed form and assumes a wild type transmembrane orientation.

Triton solubility of mycASGPR/LMP-1

A fraction of LMP-1, up to 70% in some LMP-1 expressing cell types, is insoluble in nonionic detergents (Liebowitz et al., 1987; Mann and Thorley-Lawson, 1987). LMP-1's insolubility in nonionic detergents is thought to result from an interaction with the detergent insoluble cytoskeleton (Liebowitz et al., 1987; Mann and Thorley-Lawson, 1987). Deletion analyses have identified a role for the N-terminus in rendering LMP-1 insoluble in Triton X-100 (Wang et al., 1988a,b), and have localized this contribution to LMP-1's cytoplasmic 25 amino acids (Martin and Sugden, 1991b). We explored the protein-coding requirements of the cytoplasmic N-terminus for triton-insolubility following transient expression of LMP-1 mutants in HEp2 cells (Martin and Sugden, 1991b). The relative amounts of immunoreactive LMP-1 in detergent solubilized fractions of transfected HEp2 cells are shown in Figure 3. Similar levels of LMP-1 and mycASGPR/LMP-1 were found in the triton-insoluble fraction (TI), whereas the NΔ25 deletion mutant and ASGPR/LMP-1 chimeras .4, .17, .20 were found primarily in soluble fractions (Figure 3, and not shown). Thus, the cytoplasmic N-terminus of the mycASGPR/LMP-1 chimera can effectively substitute for LMP-1's N-terminus in rendering LMP-1 triton-insoluble.

Figure 3

Cytoskeletal association of LMP-1 and mycASGPR/LMP-1. HEp2 cells electroporated with CMV-based expression vectors were fractionated as described in Materials and methods, and analysed by SDS–PAGE and Western blot with anti-LMP-1 antisera. The expressed LMP-1 protein is indicated to the left of each blot. UF, unfractionated lysate; TI, triton-insoluble fraction; W1-W2, first and second triton-soluble fractions; B958, extract of B958 cells included as a marker for LMP-1 migration. ‘Cytoskeletal association’ is defined operationally as the fraction of LMP-1 that is resistant to repeated triton extraction, and is represented by the LMP-1 protein remaining in the triton insoluble (TI) fraction

Plasma membrane patching of mycASGPR/LMP-1

LMP-1 localizes in discrete patches in the plasma membrane of transfected and infected cells (Liebowitz et al., 1986; Mann et al., 1985). LMP-1's patched phenotype correlates with transformation and cytotoxicity (Martin and Sugden, 1991b), and the N-terminus and transmembrane domains, but not the last 55 amino acids of the cytoplasmic C-terminus, are required for patching (Martin and Sugden, 1991b; Wang et al., 1988a,b). Recent studies have dissociated the patching phenotype from LMP-1's ability to activate NF-κB (Bloss et al., 1999). The role of the cytoplasmic N-terminus in plasma membrane patching was examined following expression of LMP-1 and the mycASGPR/LMP-1 chimera in 293 cells (not shown) and in the B lymphoma cell line DG75. The lyLMP-1 protein (NΔ128 deletion mutant) was included as a nonpatching LMP-1 control (Wang et al., 1988a,b). Immunofluorescence analysis of fixed and permeabilized cells revealed that, in contrast to the punctate staining pattern of LMP-1 (Figure 4a), the NΔ25 deletion mutant (not shown and Martin and Sugden, 1991b) and lyLMP-1 (Figure 4b), exhibited a diffuse membranous staining pattern. When expressed from the RSV-LTR at levels similar to those of LMP-1 in EBV-positive lymophoblastoid cell lines (Martin et al., 1993), the mycASGPR/LMP-1 chimera staining pattern was consistantly indistinguishable from characteristic LMP-1 punctate patches (Figure 4c). The patched phenotype of mycASGPR/LMP-1 was inconsistenty observed when high levels of protein were expressed (i.e. from the CMV promoter, not shown), presumably because of accumulation of nonpatching mycASGPR/LMP-1 degradation products (i.e. NΔ43; Martin and Sugden, 1991b) which would be predicted to mask the characteristic LMP-1 patches. These results demonstrate that the cytoplasmic N-terminus of the mycASGPR/LMP-1 chimera can substitute for LMP-1's N-terminus in plasma membrane patching.

Figure 4

Subcellular localization of LMP-1 and mycASGPR/LMP-1. DG75 cells transfected with RSV-based expression vectors were fixed and stained for immunofluorescence analysis using anti-LMP-1 antisera and anti-rabbit FITC conjugate as described in Materials and methods. Expressed proteins: (a), LMP-1; (b) lyLMP-1 (NΔ128 deletion mutant); (c) mycASGPR/LMP-1

Oligomerization of mycASGPR/LMP-1

The current model for LMP-1 signaling proposes that LMP-1 functions as a Tumor Necrosis Factor Receptor (TNFR) analog and signals as a constitutive oligomer (Eliopoulos and Rickinson, 1998; Eliopoulos and Young, 1998; Floettmann and Rowe, 1997; Gires et al., 1997; Hatzivassiliou et al., 1998). LMP-1 interacts with itself via a mechanism requiring the N-terminal cytoplasmic 20 amino acids and hydrophobic transmembrane domain (Gires et al., 1997). To determine if the cytoplasmic portion of the N-terminus contributes to oligomerization, we assessed the relative abilities of LMP-1 and mycASGPR/LMP-1 to oligomerize using a modification of the coimmunoprecipitation assay developed by Gires et al. (1997). The LMP-1 deletion mutant CΔ55, lacking the C-terminal 55 amino acids (Martin and Sugden, 1991b), was chosen as an oligomerization partner because LMP-1's cytoplasmic C-terminus is dispensable for oligomerization (Gires et al., 1997). In addition, CΔ55 can be resolved from full length LMP-1 by SDS–PAGE and is easily detected with our anti-LMP-1 sera (Erickson and Martin, 2000; Martin and Sugden, 1991b).

To determine the relative affinity of mycASGPR/LMP-1 for itself and for LMP-1, the C-terminus of mycASGPR/LMP-1 was truncated by 55 amino acids and tagged with the HA-1 epitope of influenza (TRmycASGPR/LMP-1.HA) (Figure 5a). This mutant is the equivalent of CΔ55, but with the heterologousN-terminus. Cells were transfected with pCMV-TRmycASGPR/LMP-1.HA and either pCMV-LMP-1 (Figure 5a,b top) or pCMV-NΔ25 (Figure 5a,b bottom). Cells were harvested 24 h later and extracts immunoprecipitated with anti-HA-1 or anti-myc antibodies. Precipitates were visualized by Western blot using anti-LMP-1 sera (Figure 5b). Anti-HA-1 immunoprecipitates contained equivalent amounts of LMP-1 and TrmycASGPR/LMP-1.HA-1 (Figure 5b, top blot). An absolute requirement for the cytoplasmic N-terminus in LMP-1 oligomerization is shown by the inability of the NΔ25 mutant to interact with TrmycASGPR/LMP-1.HA-1 chimera (Figure 5b, bottom blot). The truncated mycASGPR/LMP-1.HA-1 chimera interacted efficiently with both LMP-1 and full length mycASGPR/LMP-1 chimera as assayed by coimmunoprecipitation with the HA-1 antibody (not shown). Surprisingly, although mycASGPR/LMP-1:LMP-1 oligomers and mycASGPR/LMP-1 homo-oligomers were efficiently precipitated with the HA-1 antibody, immunoprecipitation with the myc antibody was inefficient (Figure 5b and not shown). The inefficient precipitation of mycASGPR/LMP-1 complexes with anti-myc antibodies suggests that myc antibody binding to the N-terminus of mycASGPR/LMP-1 may disrupt oligomerization, or conversely, oligomerization may decrease the efficiency of anti-myc antibody binding to the N-terminus. Nonetheless, mycASGPR/LMP-1 oligomerizes efficiently with both wild type LMP-1 and itself. An N-terminal cytoplasmic domain, presumably with positive net charge, must be present but the primary sequence is not critical. These results suggest that properly oriented transmembrane segments are the likely protein–protein interacting domain critical for the interaction of LMP-1 with itself, and that the cytoplasmic N-terminus plays a role in oligomerization via specification of wild type transmembrane topology.

Figure 5

Oligomerization of LMP-1 and mycASGPR/LMP-1. (a) Schematic of LMP-1 mutants used in coimmunoprecipitation analyses. The dotted line and asterisk in NΔ25 are meant to indicate that the majority of NΔ25 protein migrates on SDS–PAGE at the same position as the NΔ43 LMP-1 deletion mutant lacking the N-terminus and first transmembrane domain. The box in TrmycASGPR/LMP-1 represents the heterologous cytoplasmic N-terminus and the circle represents the C-terminal HA-1 tag. (b) Coimmunoprecipitation of full-length LMP-1 or NΔ25 with TrmycASGPR/LMP-1.HA-1. 293 cells were electroporated with the indicated CMV-based expression vectors, harvested 24 h post-transfection, and cell lysates were immunoprecipitated with anti-myc or anti-HA-1antibodies as shown beneath each blot. Immunoprecipitates (ip) and whole cell lysates (W) from transfected cells were resolved on 10% SDS–PAGE gels and visualized by Western blot using anti-LMP-1 antisera. The introduced LMP-1 protein is shown above each blot with a ‘+ or −’, and its migration is marked by arrows. Specificity of HA-1 and myc antibodies was confirmed by replacement of HA-1 and myc monoclonals with a matched but irrelevant monoclonal in identical immunoprecipitations (not shown)

TRAF binding to mycASGPR/LMP-1

LMP-1 activation of NF-κB and AP-1 requires TRAF and TRADD binding to CTAR sequence motifs within LMP-1's C-terminus (Brodeur et al., 1997; Devergne et al., 1996; Eliopoulos and Young, 1998; Izumi and Kieff, 1997; Kieser et al., 1997; Mosialos et al., 1995; Sandberg et al., 1997; Sylla et al., 1998). We next tested whether the mycASGPR/LMP-1 chimera retained the ability to bind to TRAF proteins. Following electroporation of 293 cells with expression vectors encoding LMP-1myc, mycASGPR/LMP-1, and HA-1 epitope-tagged TRAF3, extracts were immunoprecipitated with anti-TRAF-3 antibodies and analysed by Western blot using anti-LMP-1 antisera. LMP-1myc and mycASGPR/LMP-1 could be immunoprecipitated with TRAF3 antibodies (Figure 6), whereas the truncated lyLMP-1 variant (NΔ128) could not, despite the presence of CTAR1 and 2 domains (in preparation). LMP-1 and mycASGPR/LMP-1 were found in anti-TRAF3 immunoprecipitates from cells transfected without the TRAF3 expression vector, presumably due to interaction with the low levels of endogenous TRAF3. These results indicate that while TRAF3 binding to LMP-1's C-terminus requires an N-terminal cytoplasmic domain (presumably charged), the absolute sequence is not critical. The inability of NΔ128 to interact detectably with TRAF3 indicates that the C-terminal signaling domain is not sufficient to mediate TRAF binding in intact cells.

Figure 6

TRAF binding to LMP-1 and mycASGPR/LMP-1. 293 cells were transfected with pCMV-LMP-1myc, pCMV-mycASGPR/LMP-1 or pCMV-lyLMP-1, with or without pCMV-TRAF3/HA-1. Cells were harvested 24 h post-transfection, and extracts were immunoprecipitated with anti-TRAF3 antibodies, as described in Materials and methods. TRAF3 immunoprecipitates (ip) and whole cell lysates of transfected cells (W) were resolved by SDS–PAGE on 10% acrylamide gels and analysed by Western blot using either anti-LMP-1 antisera (LMP-1) or anti-HA-1 monoclonal antibodies (HA-1) as shown above each set of blots. The expressed LMP-1 protein is indicated above each LMP-1 blot. ‘+’ or ‘−’, the presence or absence of introduced pCMV-TRAF3/HA-1 (LMP-1 blot) or pCMV-LMP-1 (HA-1 blot) expression vectors; ‘Φ’, no introduced DNA; ‘A-D’, whole cell lysates of transfected cells expressing (a) LMP-1, (b) LMP-1myc, (c) lyLMP-1, and (d) mycASGPR/LMP-1, which were included as markers for the migration of LMP-1 and LMP-1 mutants. Arrows point to the introduced LMP-1 proteins or TRAF3/HA-1 protein

NF-κB activation by mycASGPR/LMP-1

LMP-1 is coupled to activation of cellular signaling pathways by TRAF and TRADD protein binding to its C-terminus. The NF-κB stimulating activity of LMP-1 is frequently used as a biological assay for LMP-1 signaling. Deletion of LMP-1's cytoplasmic N-terminus greatly reduces NF-κB activation (Huen et al., 1995; Mitchell and Sugden, 1995). The sequence requirements of LMP-1's cytoplasmic N-terminus for activation of cell signaling were assessed by measuring relative NF-κB activation following transient expression of LMP-1 and mycASGPR/LMP-1 in 293 cells. We chose to perform NF-κB, oligomerization, and TRAF binding assays in 293 cells because NF-κB activation in 293 cells by LMP-1 is efficient and allows accurate quantification of activity of mutants. Structure/function analyses of LMP-1 with respect to NF-κB activation, TRAF binding, and oligomerization have been carried out extensively in 293 cells, allowing comparison of mutant activities across studies (Brodeur et al., 1997; Devergene et al., 1998; Gires et al., 1997; Huen et al., 1995; Izumi and Kieff, 1997; Izumi et al., 1999; Kaye et al., 1996; Mitchell and Sugden, 1995; Sylla et al., 1998). Cells were transfected with LMP-1 expression vectors and reporter plasmids (luciferase driven by minimal fos promoter with three upstream κB binding sites (Mitchell and Sugden, 1995), and TK-lacZ to normalize luciferase activity for DNA uptake). The lyLMP-1 (NΔ128) expression vector was included as a nonfunctional LMP-1 control (Erickson and Martin, 2000; Mitchell and Sugden, 1995). Cells were harvested 48 h post-transfection and extracts were assayed for luciferase activity, β-gal activity, and LMP-1 expression. Figure 7a shows normalized luciferase activity relative to the activity induced by LMP-1. The relative levels of LMP-1 proteins expressed in each transfection were similar (Figure 7b).

Figure 7

Activation of NF-κB by LMP-1 and mycASGPR/LMP-1. (a) 293 cells electroporated with an NF-κB-responsive luciferase reporter, a TK-lacZ reporter (to monitor transfection efficiency), and the indicated CMV-driven LMP-1 expression vectors, were assayed 48 h post-transfection for luciferase/β-gal activity. Data are expressed as normalized luciferase values (relative light units) relative to activation with LMP-1, with activation by LMP-1 set at 100% (‘% Activity’). The expressed LMP-1 protein is shown in the key to the right of the graph; micrograms of DNA, amount of introduced expression vector. (b) Western analysis of expressed LMP-1 proteins. Samples from ‘A’ were also assayed for LMP-1 protein levels by Western blot using the anti-LMP-1 antisera recognizing the C-terminus. The amount of LMP-1 protein detected in each lane is representative of duplicate transfections in a single experiment. (1), lyLMP-1; (2), LMP-1; (3), mycASGPR/LMP-1. Arrows point to LMP-1 and lyLMP-1 proteins. These results are representative of five independent experiments, with the LMP-1 stimulated fold induction of luciferase/β-gal activity ranging from 15–30-fold

The mycASGPR/LMP-1 chimera activated NF-κB to levels equivalent to those activated by wild type LMP-1 whereas lyLMP-1 (NΔ128) was greatly compromised in its ability to activate NF-κB (Erickson and Martin, 2000; Huen et al., 1995; Mitchell and Sugden, 1995). NF-κB activation by mycASGPR/LMP-1 was indistinguishable from LMP-1 activation over a wide range of input DNA amounts (not shown). Thus the cytoplasmic N-terminus contributes structurally, in a sequence independent manner, to LMP-1 activation of NF-κB.

Transformation by mycASGPR/LMP-1

LMP-1 expression renders the growth of certain rodent cell lines anchorage-independent and such transformed cells are tumorigenic in nude mice (Baichwal and Sugden, 1988; Wang et al., 1985). Rodent cell transformation thus serves as a functional readout for LMP-1's biological activity and may reflect activation of signaling pathways that may not completely overlap with those involved in B cell transformation (Baichwal and Sugden, 1989; Roberts and Cooper, 1998).

We used this assay to assess the sequence requirements of LMP-1's cytoplasmic N-terminus for cellular growth regulation. Rat-1 cells were transfected with LMP-1 expression vectors and scored for focus formation 3 weeks post-transfection. There was no detectable difference in focus forming activity between LMP-1 and mycASGPR/LMP-1 in this assay (Figure 8). As expected, no transformed foci were detected in lyLMP-1 transfected cells (Baichwal and Sugden, 1989; Wang et al., 1988a) or in control vector transfectants. Equivalent amounts of LMP-1 and LMP-1 mutants were expressed following transfection (Figure 8b). Thus, although the cytoplasmic N-terminus is essential for transformation (Baichwal and Sugden, 1989; Hammerschmidt et al., 1989), this domain can be replaced functionally by a heterologous sequence with a similar net positive charge. These results demonstrate that the sequence of LMP-1's 20 residue cytoplasmic N-terminus is not critical for LMP-1's transforming properties, and suggest strongly that this domain does not contribute to transformation as a specific protein-interacting effector domain but rather serves as a structural domain.

Figure 8

Transformation of Rat-1 cells by LMP-1. Rat-1 cells were transfected with 3 μg of pRc/RSV, pRSV-LMP-1, pRSV-mycASGPR/LMP-1, or pRSV-lyLMP-1 and 24 h later either harvested for Western analysis using anti-LMP-1 antisera (b) or plated in 0.5 mg/ml G418. Three weeks post-transfection, plates were scored for foci (inset Table) and representative foci were photographed (a). The table shows two representative experiments. ‘LMP-1-1 derivative’, represents the introduced LMP-1 expression vector. Lanes in part B are: (1) B958 cell extract included as a control for migration of LMP-1, (2) pRSV-LMP-1 transfected cells, (3) pRSV-mycASGPR/LMP-1 transfected cells, (4) pRSV-lyLMP-1 transfected cells. The arrow points to the migration of the LMP-1 protein

Cytostasis induction by mycASGPR/LMP-1

LMP-1 exerts a negative growth regulatory effect on the cell (Floettmann et al., 1996; Hammerschmidt et al., 1989; Kaykas and Sugden, 2000). This activity, cytostasis, is detectable when LMP-1 is expressed in as little as twofold higher amounts than that expressed in EBV-positive lymphoblastoid cells (Kaykas and Sugden, 2000). The hydrophobic TM domain plays a critical role in LMP-1-induced cytostasis, whereas the C-terminal signaling domain is dispensable (Kaykas and Sugden, 2000). Thus, cytostasis represents a novel LMP-1 activity independent of signaling functions elicited by LMP-1's C-terminus. We compared the activity of mycASGPR/LMP-1 to that of LMP-1 in the single cell cytostasis assay modified from Kaykas and Sugden (2000). HEp2 cells were plated at clonal density 24 h post-transfection and assayed for LMP-1 expression by immunofluorescence 72 h later. The number of positive cells per colony were scored for LMP-1, mycASGPR/LMP-1 and lyLMP-1 (Figure 9). The majority of LMP-1-positive cells were either single cells or in colonies with one or two positive cells. The mycASGPR/LMP-1 chimera behaved much like LMP-1 in this assay, with the majority of positive cells scoring as either single cells or within clones with one or two positive cells. This is in contrast to lyLMP-1 transfected cells which could be found in clones with as many as 11 positive cells. Furthermore, LMP-1 and mycASGPR/LM-1 expressing cells were much larger on average, and often harbored multiple nuclei, than either untransfected neighbors or lyLMP-1 expressing cells. Similar results were obtained with the long term outgrowth assay (Hammerschmidt et al., 1989; Kaykas and Sugden, 2000) in the EBV-positive B cell line HH514 (not shown). Thus, LMP-1 and mycASGPR/LMP-1 inhibit cell proliferation whereas lyLMP-1 does not. Consistent with results of other LMP-1 functional assays, these results suggest the cytoplasmic N-terminus does not appear to contribute to cytostasis directly, but rather functions in a structural capacity. Together with results of Kaykas and Sugden (2000), these results strongly suggest that LMP-1's hydrophobic transmembrane domain, independent of both cytoplasmic termini, is sufficient to induce cytostasis so long as it is properly anchored in the plasma membrane.

Figure 9

LMP-1 Induced Cytostasis. HEp2 cells were transfected with pCMV-LMP-1, pCMV-mycASGPR/LMP-1 or pCMV-lyLMP-1 and plated on coverslips at clonal density 24 h later. Three days post-plating, cells were fixed and stained with anti-LMP-1 antisera as described in Materials and methods. All colonies on a given coverslip were scored for the number of positive cells (LMP-1 immunoreactive) and data are expressed as # positive cells/colony. ‘1*’ represents the number of single LMP-1 positive cells (isolated positive cells, ‘colony’ of one). A total of 500 colonies were scored for each transfection. Shown below the graph are representative colonies from each transfection. These results are from a single experiment but are representative of three separate experiments

Turnover of mycASGPR/LMP-1

The LMP-1 protein has a short half life in EBV-positive lymphoblastoid cell lines and LMP-1-transformed rodent cells (Baichwal and Sugden, 1987). LMP-1's rapid turnover correlates with many of its biochemical and functional properties (Martin and Sugden, 1991b). Importantly, the properties of LMP-1's turnover resemble turnover/downregulation of activated cell surface receptors (Martin and Sugden, 1991a). These properties of LMP-1's turnover are consistent with LMP-1's proven function as a constitutively active cell surface receptor analog. Because LMP-1 is constitutively active, mechanisms must exist within the cell to regulate LMP-1 activity. Enhanced turnover (i.e. downregulation) triggered by constitutively activated signaling pathways could be one such mechanism (Martin and Sugden, 1991a). We compared the half life of mycASGPR/LMP-1 to that of LMP-1 (Figure 10). We found, like other functional LMP-1 variants tested previously (Martin and Sugden, 1991b), that mycASGPR/LMP-1 has a half life indistinguishable from LMP-1's (compare to nonfunctional NΔ25 deletion mutant). We have performed these experiments in Balb/c 3T3 and HEp2 cells, with similar results (not shown). Although the cytoplasmic N-terminus contributes to turnover (i.e. NΔ25 does not turn over as rapidly as does LMP-1), our results suggest LMP-1's turnover does not depend on the absolute sequence of LMP-1's cytoplasmic N-terminus. This is intriguing given the recent results suggesting an N-terminal sequence requirement for LMP-1's ubiquitination and turnover (Aviel et al., 2000) (see Discussion). We propose that the contribution of the N-terminus is structural, providing the basis for proper transmembrane topology and therefore rendering LMP-1 constitutively active and subject to downregulation at the level of turnover.

Figure 10

Turnover of LMP-1 and mycASGPR/LMP-1. HH514 cells were transfected with pCMV-driven expression vectors for LMP-1, mycASGPR/LMP-1 and NΔ25, and analysed by pulse/chase 48 h post-transfection as described in Materials and methods. Data are plotted as ‘log pixel density vs time of chase’. Half lives calculated from linear regression were 7 h for LMP-1, 6.9 h for mycASGPR/LMP-1 and >20 h for NΔ25 (the t1/2 for NΔ25 could not be accurately determined because the calculated t1/2 (26 h) was longer than the longest time point in the experiment)


Deletion analyses of the LMP-1 open reading frame have identified the cytoplasmic N-terminus as an essential domain for transformation, NF-κB activation, subcellular localization and cytoskeletal association (Baichwal and Sugden, 1989; Hammerschmidt et al., 1989; Huen et al., 1995; Liebowitz et al., 1992; Martin and Sugden, 1991b; Mitchell and Sugden, 1995; Moorthy and Thorley-Lawson, 1993). The possible contribution of the cytoplasmic N-terminus as an effector domain has been previously investigated by generation of LMP-1 mutants in which three potential sequence motifs were individually deleted. These mutants were generated in the context of the EBV genome, and recombinants assayed for transformation efficiency in human B-cells (Izumi et al., 1994). Deletion of the EHDLER sequence (amino acids 2–7) or GPPLSSS (amino acids 18–24) does not diminish the transforming efficiency of recombinant EBV, whereas deletion of the sequence encompassing the ERGPPGPRRPPR motif (amino acids 6–17 or 6–24) results in a 10–600-fold decrease in transformation efficiency (Izumi et al., 1994). Interestingly, the amino-termini of Δ6–17 and Δ6–24 have a net charge of −0.5 as compared to +2.5 and +4.0 for Δ18–24 and Δ2–7, respectively (the net charge of the wild type cytoplasmic N-terminus is +2.5). These results suggest that the N-terminus functions as a structural domain as a result of its net positive charge (Izumi et al., 1994). The possibility, however, that the proline/arginine rich sequence encoded by residues 7–20 contributes to function as an effector domain could not be ruled out due to the fact that it overlaps the positively charged, basic residues and neither motif can be selectively deleted. Mitchell and Sugden (1995) subsequently reported that the pro/arg rich region of the N-terminus (residues 12–20) was dispensible for LMP-1's activation of NF-κB and suggested a structural role for this domain (Bloss et al., 1999).

To address experimentally whether the cytoplasmic N-terminus of LMP-1 serves as a structural domain, we have generated an LMP-1 replacement mutant in which LMP-1's first 20 amino acids are replaced with an unrelated domain possessing a similar charge distribution across the cytoplasmic N-terminus (N-terminus acidic, C-terminus basic). Our goal was to create a chimera likely to retain wild type topological information (i.e. net positive charge across the first transmembrane domain) thereby retaining structural determinants of the N-terminus. This chimera would then allow us to dissociate potential N-terminal effector and structural functions. We found, using proteolytic analysis, that the mycASGPR/LMP-1 chimera was indistinguishable from LMP-1 in its plasma membrane orientation (Figure 2). In contrast, cleavage products generated from NΔ25 and the ASGPR.4, .17 and .20 chimeras were distinct from those obtained from wild type LMP-1. ASGPR.4, .17 and .20 chimeras, despite having similar positive charge distributions as mycASGPR/LMP-1 across the first transmembrane domain, lack the LSSSL sequence at the N-terminus of transmembrane domain 1 and are inserted improperly in the plasma membrane (Figure 2, and data not shown). Interestingly, the termini of the improperly inserted ASGPR/LMP-1 chimeras were intracellular. These results indicate that a charged N-terminus is not sufficient for proper membrane anchoring and suggest that the ability of the first transmembrane domain to span the lipid bilayer may also be a contributing factor in determining transmembrane orientation. Furthermore, wild type positioning of LMP-1's N- and C-termini within the cell does not guarantee that the membrane spanning domains will be inserted properly in the membrane. In retrospect, structure/function analyses of LMP-1 using the NΔ25 deletion mutant (and presumably N-terminal deletion mutants such as NΔ43 as well) are difficult to interpret because loss of function apparently reflects aberrant insertion into the plasma membrane (Figure 2). The topological mapping shown in Figure 2, particularly with respect to NΔ25, emphasizes the inherent difficulties in interpreting results of deletion analyses of polytopic membrane proteins.

LMP-1 and mycASGPR/LMP-1 exhibited similar properties when assayed for triton insolubility and plasma membrane patching. LMP-1's insolubility in nonionic detergents may reflect an interaction with the triton-insoluble cytoskeleton (Liebowitz et al., 1987; Mann and Thorley-Lawson, 1987), or with other triton-insoluble structures within the cell such as glycosphingolipid-rich complexes (Clausse et al., 1997). LMP-1's triton-insolubility requires the cytoplasmic N-terminus and correlates with function (Martin and Sugden, 1991b). As shown previously (Martin and Sugden, 1991b), deletion of the first 25 amino acids rendered LMP-1 triton-soluble (Figure 3). Like NΔ25, the ASGPR/LMP-1 chimeras .4, .17, and .20 were soluble in triton (not shown). However, replacement of residues 1–20 with the mycASGPR sequence effectively restored LMP-1's triton-insolubility. Similarly, LMP-1 and mycASGPR/LMP-1 were indistinguishable in their plasma membrane localization (Figure 4). Like LMP-1, mycASGPR/LMP-1 exhibited a patched, punctate staining pattern distinct from that of the nonpatching NΔ128 (lyLMP-1) mutant. These results indicate that the absolute sequence of the cytoplasmic N-terminus is not critical, and that this domain contributes structural information, to LMP-1's observed insolubility in nonionic detergents and plasma membrane patching. Previous studies have shown that LMP-1's cytoplasmic C-terminus is not required for patching or triton insolubility (Martin and Sugden, 1991b). Our results suggest a critical role for properly oriented transmembrane domains, independent of the presence of the cytoplasmic C-terminus and absolute sequence of N-terminus, in patching and triton insolubility.

The key regions of the C-terminus necessary for signaling, CTAR-1 and CTAR-2, function by binding cellular TRAF and TRADD signaling proteins. Oligomerization of CTAR domains greatly increases LMP-1's interaction with TRAF proteins. Consistent with its wild type signaling activity, the mycASGPR/LMP-1 chimera retained the ability to oligomerize (Figure 5) and bind to TRAF proteins (Figure 6). The interference with oligomerization by the N-terminally directed myc antibody suggests that domains in physical proximity to the cytoplasmic N-terminus may constitute a portion of the binding interface of the LMP-1 oligomer. LMP-1's C-terminus is dispensible for homo-oligomerization (Gires et al., 1997), and while the cytoplasmic N-terminus is required, its contribution is secondary in that it functions to position the essential hydrophobic transmembrane domain in the membrane. Thus, our results show that correctly oriented membrane spanning segments, independent of intracellular termini, are the primary protein–protein interacting sequences involved in oligomerization. These findings support the recently proposed model in which LMP-1's hydrophobic membrane spanning domain is proposed to mediate hetero-oligomerization with cellular membrane proteins to trigger LMP-1-induced cytostasis (Kaykas and Sugden, 2000).

Deletion of LMP-1's cytoplasmic N-terminus (as in NΔ25) greatly reduces NF-κB activation by LMP-1 (Huen et al., 1995; Mitchell and Sugden, 1995), presumably as a result of improper transmembrane orientation (Figure 2). Replacement of residues 1–20 with a domain of like charge, but different primary sequence, was sufficient to restore the ability of LMP-1 to activate NF-κB (Figure 7), whereas the ASGPR.4, .17, and .20 chimeras, which lack juxtamembrane residues 21–25, were indistinguishable in function from NΔ25 (not shown). Mitchell and Sudgen (1995) have shown that the pro/arg rich subregion of the cytoplasmic N-terminus (residues 12–20) is not required for NF-κB activation and suggest a role for the N-terminus in stabilization of the adjacent transmembrane domain. Consistent with their results, we find that the cytoplasmic N-terminus does not function as an NF-κB-activating domain per se. Rather, we show here that the entire cytoplasmic N-terminus (residues 1–20) can be replaced functionally with an unrelated, but positively charged sequence of similar size, so long as the first transmembrane domain is intact. Importantly, the ability of the mycASGPR/LMP-1 replacement mutant to induce focus formation in Rat-1 cells in a manner indistinguishable from LMP-1 argues strongly that no sequence in LMP-1's cytoplasmic N-terminus encodes an effector domain that contributes to LMP-1's transforming activity.

Induction of cytostasis by mycASGPR/LMP-1 (Figure 9) supports the results of Kaykas and Sugden (2000) showing LMP-1's N-terminal and hydrophobic transmembrane domain are sufficient to induce cytostasis. The activity of mycASGPR/LMP-1 in this assay support strongly the notion that the hydrophobic domain is sufficient, and that no specific sequence within the first 20 amino acids of LMP-1 is required, for induction of cytostasis. NΔ25's lack of cytostatic activity (Kaykas and Sugden, 2000) indicates that expression of LMP-1's transmembrane domain is not sufficient for activity, but rather its proper insertion into the membrane is critical.

Previous studies have demonstrated a correlation between LMP-1's rapid turnover and its transforming/cytotoxic activity (Martin and Sugden, 1991b). LMP-1's turnover is preceded by internalization, and once internalized, LMP-1 is rapidly degraded (Martin and Sugden, 1991a). Like ligand-activated receptors (i.e. epidermal growth factor and insulin receptors (Gross et al., 1983; Knutson et al., 1985), LMP-1's turnover, but not internalization, requires ongoing protein synthesis (Martin and Sugden, 1991a). The properties of LMP-1's turnover, and the correlation between constitutive activity and turnover (Martin and Sugden, 1991b), suggest LMP-1's turnover is triggered by activated signaling and results in LMP-1's downregulation. Thus, turnover may be a cellular mechanism aimed at dampening LMP-1's constitutive growth regulatory signaling. The rapid turnover of mycASGPR/LMP-1 is consistent with this model. Like LMP-1, mycASGPR/LMP-1 activates pathways culminating in NF-κB activation, rodent cell transformation and cytostasis and thus would be subject to this downregulatory mechanism. The rapid turnover of mycASGPR/LMP-1 in the cell lines reported here (HH514, HEp2 and Balb/c3T3) is particularly intriguing given that insertion of the myc tag at LMP-1's N-terminus has been recently reported to significantly stabilize LMP-1 (Aviel et al., 2000). Aviel et al. (2000) suggest the increased stability of N-terminally tagged LMP-1 (and N-terminal deletion mutants) results from alteration or removal of a recognition/targeting motif for ubiquitination of the α-amino group at LMP-1's N-terminal residue. The N-terminus of the myc-tagged LMP-1 variant reported by Aviel et al. (2000) is presumably identical in sequence to mycASGPR/LMP-1 up to residue 10. The mycASGPR/LMP-1 chimera, based on the model presented in Aviel et al. (2000) would be predicted to turn over with a significantly longer half life than does LMP-1. These apparently conflicting results may not be directly comparable given the differences in cell lines and methods for determintion of turnover. However, it is conceivable that LMP-1 turnover is the sum of distinct regulatory mechanisms, one of which may be downregulation in response to constitutive signaling as we suggest, and another may be N-terminal ubiquitination and proteosomal degradation as suggested by Aveil et al., 2000.

Our results provide the first experimental evidence that the primary function of LMP-1's cytoplasmic N-terminus is to ensure proper orientation of the transmembrane domains. Furthermore, our findings demonstrate that properly oriented transmembrane segments serve as the primary domain mediating LMP-1 oligomerization and cell surface localization, both of which are critical for LMP-1 coupling to downstream effectors. This function of the hydrophobic transmembrane domain is independent of LMP-1's intracellular N- and C-termini.

Materials and methods

Generation of chimeric LMP-1 expression vectors

The cytoplasmic N-termini of all chimeras were designed to have a positive net charge (as does LMP-1's N-terminus), and to have a negative charge difference across the first transmembrane domain (Δ(C-N)). A negative charge difference correlates with an Ncyt/Cexo orientation across the membrane (Hartmann et al., 1989). The pcDNA3-based plasmid pCMV-mycASGPR/LMP-1 encodes an LMP-1 chimera in which the first 20 amino acids of LMP-1 (MEHDLERGPPGPRRPPRGPP) are replaced with the sequence MEQKLISEEDL RK SDHHQLRKGPR: myc tag, RK added to bring the net charge of the N-terminus to +2.0, residues 16–25 of the asialoglycoprotein receptor H1 (ASGPR-H1) N-terminus (Spiess et al., 1985).


Underlined sequences are from the N-terminus of the ASGPR H1 cDNA (gift of H Lodish) and were inserted into pCMV-LMP-1 (modified from pCMV-BNLF-1 (Hammerschmidt et al., 1989) to encode a unique KpnI site upstream of the initiating methionine for LMP-1). The resulting LMP-1 chimeras encode residues 1–40 (.20), 1–24 (.4 and .17) from the ASGPR H1 in place of the first 25 amino acids of the wild type LMP-1 N-terminal sequences described above. ASGPR/LMP-1.4 differs from ASGPR/LMP-1.17 in that it encodes the amino acids DYKD from the FLAG epitope (after the initiating methionine). ASGPR/LMP-1.20, .17, and .4 differ from mycASGPR/LMP-1 in that they lack residues 21–25 (LSSSL) from the N-terminal region of LMP-1's first transmembrane domain (these sequences were lost in cloning). The primary sequence of all chimeras has been confirmed by DNA sequencing.

Other plasmids

pCMV-LMP-1 and pCMV-lyLMP-1 are pcDNA3-based vectors encoding LMP-1 or the N-terminally truncated LMP-1 protein (referred to as either lyLMP-1 or NΔ128), respectively. pRSV-LMP-1, pRSV-lyLMP-1, and pRSV-mycASGPR/LMP-1 are pRc-RSV-based vectors (Invitrogen) encoding full length LMP-1, lyLMP-1 (NΔ128), or mycASGPR/LMP-1 proteins, respectively. pCMV-LMP-1myc was constructed from pCMV-LMP-1 and encodes the 10 amino acid myc epitope at LMP-1's C-terminus. pCMV-NΔ25 encodes the LMP-1 deletion mutant NΔ25, which lacks the first 25 codons from the N-terminus (Martin and Sugden, 1991b). pCMV-TrmycASGPR/LMP-1.HA-1 encodes a C-terminal deletion (CΔ55) of the mycASGPR/LMP-1 chimera with an HA-1 epitope tag (YPYDVPDYA) at the C-terminus. p1242 is a luciferase reporter plasmid (Mitchell and Sugden, 1995) encoding the luciferase gene driven by the minimal fos promoter with three upstream κB binding sites from the MHC class I gene (gift of B Sugden). TK-lacZ (gift of S Konieczny) encodes the lac Z gene expressed from the thymidine kinase promoter). pCMV-TRAF3/HA-1 encodes HA-1 tagged human TRAF3, and was a gift from B Sugden (Sandberg et al., 1997).


Anti-LMP-1 antisera is an affinity-purified rabbit polyclonal sera raised against LMP-1's carboxy-terminus (residues 188–352) fused to glutathione S-transferase (Erickson and Martin, 2000). Monoclonal antibodies recognizing myc (9E10) or HA-1 (F-7) epitopes, and anti-TRAF3 polyclonal sera (H-20) were from Santa Cruz Biochemicals. Monoclonal anti-HA-1 antibody HA-11 was from Babco (gift of B Sugden).

Cells and transfections

293 is an adherent human embryonic kidney carcinoma cell line grown in DMEM/10% fetal bovine serum; HEp2 is an adherent human epithelial cell line grown in DMEM/10% calf serum. DG75 is a human EBV-negative B lymphoma cell line. HH514 is derived from the P3HR1 clone of the Jijoye BL cell line (Rabson et al., 1983) and was grown in RPMI/10% calf serum. Cells were either electroporated using a Biorad Gene Pulser and harvested 1–2 days post-transfection, or transfected in subconfluent 6-well plates using Lipofectamine (Gibco) according to the manufacturer's instructions and harvested 24 h post-transfection.

Western analysis

Transfected cells were harvested and lysed in 4×SDS-sample buffer, boiled, and resolved on 10% acrylamide gels. Proteins were transferred to Immobilon (Millipore), and stained with the appropriate antisera as follows: Blots were blocked in PBS/1% milk/0.05% Tween 20, and incubated with primary antibody (1 : 5000 anti-LMP-1 antisera or 1 : 500 anti-HA-1 monoclonal HA-11, and secondary antibody (1 : 4000 horse radish peroxidase conjugated anti-rabbit antibody, 1 : 10 000 anti-mouse HRP-conjugated antibody or 1 : 1000 anti-rabbit alkaline phosphatase conjugated antibody (Promega)). Blots were visualized by ECL according to manufacturers instructions (Amersham) or with nitro-blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Sigma).

Intact cell proteolysis

293 cells electroporated with CMV-based expression vectors were washed with PBS, resuspended in PBS, PBS/Trypsin (9 units/ml; Worthington Biochemicals), chymotrypsin (5 units/ml; Sigma) or pronase (1 unit/ml; Sigma) and rotated at 22°C for 10 min. Control experiments (not shown) demonstrated that these conditions were sufficient to cleave accessible protease sites in LMP-1, but cause minimal cell lysis. Following proteolysis, intact cells were lysed in 2×SDS-sample buffer (final concentration), sonicated and analysed by SDS–PAGE and Western blot as described above.

Localization of epitope-tagged N-terminus in permeabilized vs impermeabilized cells

DG75 cells electroporated with CMV-based expression vectors were washed in PBS, and spotted onto slides at a concentration of 105 cells/20 μl. When dry, slides were fixed in either 4% paraformaldehyde/1% Triton X-100 (permeable) or 4% paraformaldehyde (impermeable). Slides were blocked in PBS/1% calf serum and stained with either 1 : 50 anti-LMP-1 antisera or 1 : 200 anti-myc monoclonal (9E10, Santa Cruz), followed by 1 : 50 anti-rabbit FITC conjugated secondary antibody (Sigma) or 1 : 200 anti-mouse FITC conjugated secondary, respectively. Stained cells were visualized under a Nikon Eclipse E800 fluorescent microscope (40×magnification) and the percentage of positive cells determined.

Cytoskeletal association

HEp2 cells electroporated with CMV-based expression vectors were washed 3×with PBS, and extracted with 0.5 ml of a buffered 1% Triton X-100 solution (Martin and Sugden, 1991b) for 3 min with rocking every 30 s (Triton wash one, W1); the plate was then extracted again with the same buffer for 1 min (Triton wash two, W2). Triton washes one and two were diluted 1 : 1 with 4×SDS-sample buffer. The Triton-insoluble material (TI) left on the plate was solubilized in 0.5 ml 1×RIPA buffer and diluted 1 : 1 in 4×SDS-sample buffer. An identical plate (unfractionated lysate, UF) was solubilized with 0.5 ml 1×RIPA buffer and diluted 1 : 1 in 4×SDS-sample buffer. All samples were sonicated and resolved on SDS–PAGE as described above. The same number of cell equivalents were loaded on each lane.

Plasma membrane patching

DG75 cells (5×106 cells) electroporated with 1.2 μg of RSV-driven LMP-1 expression vectors were washed in PBS, dried on slides and fixed in acetone/methanol (1 : 1) for 20 min at −20°C. Slides were blocked in PBS/1% calf serum, and stained with 1 : 50 anti-LMP-1 antisera followed by 1 : 40 anti-rabbit FITC conjugated secondary antibody (Sigma). Slides were visualized under a Nikon Eclipse E800 fluorescent microscope (100×magnification), and images were captured with a Cooke SensiCam Digital camera using SlideBook software.

Oligomerization assay

LMP-1 homo-oligomerization was assayed by coimmunoprecipitation using the method described in Gires et al. (1997), with the following modifications. 293 cells transfected with CMV-based expression vectors were harvested 24 h post-transfection, lysed in hypotonic lysis buffer (10 mM HEPES, pH 7.9/0.5 mM KCl/0.5 mM MgCl2/0.1 mM EGTA/0.5 mM DTT) at 25°C for 30 min, and triturated through a 27 g needle. The supernatant obtained following microfugation was discarded, and the pellet resuspended in TNA lysis buffer (50 mM Tris, pH 7.4/0.15 mM NaCl/1% NP40/5 mM EDTA, pH 8) by vortexing and incubation on ice for 15 min. Insoluble material was removed by microcentrifugation, and the supernatant was precleared by incubation with Protein-G agarose (Boehringer Mannheim) for 90 min at 4°C. Anti-LMP-1, anti-myc (9E10) or anti-HA-1 (F-7) antibody was added to supernatant (1 : 100 dilution) and incubated at 4°C with rocking for 45 min. Immunoprecipitates were recovered with Protein-G agarose (Boehringer), and beads were washed four times in 1×RIPA buffer. The final pellet was resuspended and boiled in 4×SDS-sample buffer. Extracts were resolved by SDS–PAGE and transferred to Immobilon. Blots were stained with anti-LMP-1 antisera as described above and in legends.

TRAF binding assay

293 cells electroporated with CMV-based expression vectors were harvested and lysed essentially as described below for the oligomerization assay except that 1×Complete Protease Inhibitor (Boehringer Mannheim) was added to the hypotonic lysis buffer, TNA lysis buffer and RIPA buffer for the final washes. Anti-TRAF3 antibody H-20 was used at a dilution of 1 : 50.

Activation of NF-κB

Subconfluent 293 cells transfected in triplicate by electroporation with 1.0 μg p1242 (luciferase gene under control of 3 κB binding sites upstream of minimal fos promoter), 1.0 μg TK-lacZ (to normalize transfection efficiency), and 1.35 μg CMV-promoter driven LMP-1 expression vectors. Forty-eight hours following transfection, cells were harvested and each extract was assayed in duplicate for NF-κB activity using the Dual Light Assay from Tropix. Luciferase values were averaged for each sample and normalized with averaged β-gal values to yield relative light units (RLU). The RLU values were averaged for each set of duplicate transfections, and data expressed relative to LMP-1-stimulated activity. Data shown are representative of five independent experiments. Samples were also assayed for LMP-1 expression by Western analysis, as described above.

Rodent cell transformation assay

Rat-1 cells (5×106) were transfected with 3 μg of DNA using Lipofectamine according to the manufacturer's instructions. Twenty-four hours post-transfection, cells were either harvested for Western blot (described above) or treated with 0.5 mg/ml G418. Three weeks later plates were scored for foci and photographed under phase microscopy at 10×magnification.

HEp2 cytostasis assay

This assay is modified from that described by Kaykas and Sugden (2000) as follows. HEp2 cells were transfected by lipofection and plated the next day at clonal density on coverslips. Cells were fixed 3 days after plating and coverslips stained with anti-LMP-1 antisera as described above. All colonies on a given coverslip were scored for number of positive (LMP-1 immunoreactive) cells and expressed as # of positive cells/colony. A total of 500 colonies were scored for each transfection.

Pulse/chase analysis

HH514 cells were transfected by electroporation with 10 μg DNA and grown for 48 h. Cells (5×106/ml) were incubated in 2 ml methionine-free RPMI/10% dialyzed calf serum/50 mM HEPES, pH 7.4 for 30 min at 37°C, Tran[35S]-label (methionine/cysteine) was added (concentration of 0.5 mCi/ml, 1175 Ci/mmol; ICN) and the incubation continued for 1 h (pulse). Cells were collected by centrifugation, washed three times with PBS and either stored as frozen cell pellets (T=0) or incubated for up to 24 h in RPMI/10% calf (chase). At the time of harvest, cells were collected by centrifugation, washed in PBS and stored as frozen cell pellets until completion of the time course. Cell pellets were analysed by immunoprecipitation as described previously (Erickson and Martin, 2000), using the affinity purifed anti-LMP-1 antiserum that recognizes LMP-1's C-terminus. Immunoprecipitates were resolved on 10% SDS–PAGE gels. Gels were fixed in 10% acetic acid/30% methanol for 60 min, soaked in Amplify (Amersham) for 30 min and dried under vacuum at 80°C. Dried gels were exposed to preflashed Hyperfilm MP, with an intensifying screen, at −80°C. The t1/2 values were determined from linear exposures as described (Martin and Sugden, 1991b).


  1. Aviel S, Winberg G, Massucci M, Ciechanover A . 2000 J. Biol. Chem. 275: 23491–23499

  2. Baichwal VR, Sugden B . 1987 J. Virol. 61: 866–875

  3. Baichwal VR, Sugden B . 1988 Oncogene 2: 461–467

  4. Baichwal VR, Sugden B . 1989 Oncogene 4: 67–74

  5. Bankier AT, Deininger PL, Satchwell SC, Baer R, Farrell PJ, Barrell BG . 1983 Mol. Biol. Med. 1: 425–445

  6. Beltzer JP, Fiedler K, Fuhrer C, Geffen I, Handschin C, Wessels HP, Spiess M . 1991 J. Biol. Chem. 266: 973–978

  7. Bloss T, Kaykas A, Sugden B . 1999 J. Gen. Virol. 80: 3227–3232

  8. Bonifacino J, Lippincott-Schwartz J . 1991 Curr. Opin. Cell. Biol. 3: 592–600

  9. Bonifacino J, Weissman A . 1998 Annual Review of Cell. and Dev. Biol. 14: 19–57

  10. Brodeur SR, Cheng G, Baltimore D, Thorley-Lawson DA . 1997 J. Biol. Chem. 272: 1977–19784

  11. Clausse B, Fizazi K, Walczak V, Tetaud C, Wiels J, Tursz T, Busson P . 1997 Virology 228: 285–293

  12. Devergene O, Hatzivassiliou E, Izumi KM, Kaye K, Kleijnen MF, Kieff E, Mosialos G . 1998 J. Virol. 72: 7900–7908

  13. Devergne O, Hatzivassiliou E, Izumi K, Kaye K, Kleijnen M, Kieff E, Mosialos G . 1996 Mol. Cell. Biol. 16: 7098–7108

  14. Eliopoulos AG, Blake SS, Floettmann JE, Rowe M, Young LS . 1999 J. Virol. 73: 1023–1035

  15. Eliopoulos AG, Rickinson AB . 1998 Curr. Biol. 8: R196–R198

  16. Eliopoulos AG, Young LS . 1998 Oncogene 16: 1713–1742

  17. Erickson KD, Martin JM . 2000 J. Virol. 74: 1057–1060

  18. Fahreus R, Fu HL, Ernberg I, Finke J, Rowe M, Klein G, Falk K, Nilsson E, Yadav M, Busson P, Tursz T, Kallin B . 1988 Int. J. Cancer 42: 329–338

  19. Fennewald S, van SV, Kieff E . 1984 J. Virol. 51: 411–419

  20. Floettmann JE, Rowe M . 1997 Oncogene 15: 1851–1858

  21. Floettmann JE, Ward K, Rickinson AB, Rowe M . 1996 Virology 223: 29–40

  22. Fries KL, Miller WE, Raab-Traub N . 1996 J. Virol. 70: 8653–8659

  23. Gires O, Zimber-Strobl U, Gonnella R, Ueffing M, Marschall G, Zeidler R, Pich D, Hammerschmidt W . 1997 EMBO J. 16: 6131–6140

  24. Gratama JW, Zutter MM, Minarovits J, Oosterveer MA, Thomas ED, Klein G, Ernberg I . 1991 Int. J. Cancer 47: 188–192

  25. Gross JL, Krupp MN, Rifkin DB, Lane MD . 1983 Proc. Natl. Acad. Sci. USA 80: 2276–2280

  26. Haigler HT, Ash JF, Singer SJ, Cohen C . 1978 Proc. Natl. Acad. Sci. USA 75: 5025–5029

  27. Hammerschmidt W, Sugden B, Baichwal VR . 1989 J. Virol. 63: 2469–2475

  28. Hammerskjold M, Simurda M . 1992 J. Virol. 66: 6496–6501

  29. Hartmann E, Rapoport TA, Lodish HF . 1989 Proc. Natl. Acad. Sci. USA 86: 5786–5790

  30. Hatzivassiliou E, Miller WE, Raab-Traub N, Kieff E, Mosialos G . 1998 J. Immunol. 160: 1116–1121

  31. Hedge R, Lingappa V . 1997 Cell 28: 575–582

  32. Hedge R, Lingappa V . 1999 Trends Cell. Biol. 9: 132–137

  33. Huen DS, Henderson SA, Croom-Carter D, Rowe M . 1995 Oncogene 10: 549–560

  34. Izumi KM, Kaye KM, Kieff E . 1994 J. Virol. 68: 4369–4376

  35. Izumi KM, Kaye KM, Kieff ED . 1997 Proc. Natl. Acad. Sci. USA 94: 1447–1452

  36. Izumi KM, Kieff ED . 1997 Proc. Natl. Acad. Sci. USA 94: 12592–12597

  37. Izumi KM, McFarland EC, Ting AT, Riley EA, Seed B, Kieff ED . 1999 Mol. Cell Biol. 19: 5759–5767

  38. Kaye K, Izumi K, Mosialos G, Kieff E . 1995 J. Virol. 69: 675–683

  39. Kaye KM, Devergne O, Harada JN, Izumi KM, Yalamanchili R, Kieff E, Mosialos G . 1996 Proc. Natl. Acad. Sci. USA 93: 11085–11090

  40. Kaye KM, Izumi KM, Kieff E . 1993 Proc. Natl. Acad. Sci. USA 90: 9150–9154

  41. Kaykas A, Sugden B . 2000 Oncogene 19: 1400–1410

  42. Kieser A, Kaiser C, Hammerschmidt W . 1999 EMBO J. 18: 2511–2521

  43. Kieser A, Kilger E, Gires O, Ueffing M, Kolch W, Hammerschmidt W . 1997 EMBO J. 16: 6478–6485

  44. Knutson VP, Ronnett GV, Lane MD . 1985 J. Biol. Chem. 260: 14180–14188

  45. Laherty CD, Hu HM, Opipari AW, Wang F, Dixit VM . 1992 J. Biol. Chem. 34: 24157–24160

  46. Liebowitz D, Kopan R, Fuchs E, Sample J, Kieff E . 1987 Mol. Cell. Biol. 7: 2299–2308

  47. Liebowitz D, Mannick J, Takada K, Kieff E . 1992 J. Virol. 66: 4612–4616

  48. Liebowitz D, Wang D, Kieff E . 1986 J. Virol. 58: 233–237

  49. Lipp J, Dobberstein B . 1986 J. Cell. Biol. 102: 2169–2175

  50. Mann KP, Staunton D, Thorley-Lawson D . 1985 J. Virol. 55: 710–720

  51. Mann KP, Thorley-Lawson D . 1987 J. Virol. 61: 2100–2108

  52. Martin JM, Sugden B . 1991a Cell Growth and Differentiation 2: 653–660

  53. Martin JM, Sugden B . 1991b J. Virol. 65: 3246–3258

  54. Martin JM, Veis D, Korsmeyer SJ, Sugden B . 1993 J. Virol. 67: 5269–5278

  55. Miller G . 1985 Epstein-Barr Virus, 1 edn. Miller G (ed.) Raven Press, New York

  56. Miller WE, Cheshire JL, Raab-Traub N . 1998 Mol. Cell. Biol. 18: 2835–2844

  57. Mitchell T, Sugden B . 1995 J. Virol. 69: 2968–2976

  58. Moorthy RK, Thorley-Lawson DA . 1993 J. Virol. 67: 1638–1646

  59. Mosialos G, Birkenbach M, Yalamanchili R, Van Arsdale T, Ware C, Kieff E . 1995 Cell. 80: 389–399

  60. Paine E, Scheinman RI, Baldwin Jr. AS, Raab-Traub N . 1995 J. Virol. 69: 4572–4576

  61. Pallesen G, Hamilton-Dutolt SJ, Rowe M, Young IS . 1991 Lancet 337: 320–322

  62. Parks GD, Lamb RA . 1991 Cell. 64: 777–787

  63. Rabson M, Heston L, Miller G . 1983 Proc. Natl. Acad. Sci. USA 80: 2762–2766

  64. Roberts ML, Cooper NR . 1998 Virology 240: 93–99

  65. Sandberg M, Hammerschmidt W, Sugden B . 1997 J. Virol. 71: 4649–4656

  66. Schlessinger J, Shechter Y, Willingham MC, Pastan I . 1978 Proc. Natl. Acad. Sci. USA 75: 2659–2663

  67. Schmid SR, Spiess M . 1988 J. Biol. Chem. 263: 16886–16891

  68. Spiess M, Lodish HF . 1986 Cell. 44: 177–185

  69. Spiess M, Schwartz AL, Lodish HF . 1985 J. Biol. Chem. 260: 1979–1982

  70. Sylla BS, Hung SC, Davidson DM, Hatzivassiliou E, Malinin NL, Wallach D, Gilmore TD, Kieff E, Mosialos G . 1998 Proc. Natl. Acad. Sci. USA 95: 10106–10111

  71. Thomas JA, Hotchin NA, Allday MJ, Amlot P, Rose M, Yacoub M, Crawford DH . 1990 Transplantation 49: 944–953

  72. Wang D, Liebowitz D, Kieff E . 1985 Cell 43: 831–840

  73. Wang D, Liebowitz D, Kieff E . 1988a J. Virol. 62: 2337–2346

  74. Wang D, Liebowitz D, Wang F, Gregory C, Rickinson A, Larson R, Springer T, Kieff E . 1988b J. Virol. 62: 4173–4184

  75. Yarden Y, Ullrich A . 1988 Annu. Rev. Biochem. 57: 443

  76. Young L, Alfieri C, Hennessy K, Evans H, O'Hara C, Anderson KC, Ritz J, Shapiro RS, Rickinson A, Kieff E, Cohen JI . 1989 N. Engl. J. Med. 321: 1080–1085

Download references


The TRAF expression plasmid and anti-HA antibody HA-11 were generous gifts from Bill Sugden. The TK-lacZ reporter was the gift of Brad Olwin. This work was supported by NIH CA-64610 and AI-01537 grants to JM Martin.

Author information

Correspondence to Jennifer M Martin.

Rights and permissions

Reprints and Permissions

About this article


  • LMP-1
  • EBV
  • transmembrane topology
  • signaling

Further reading