Accumulating evidence indicates that the reduction and rearrangement of disulfide bonds constitutes a mechanism controlling protein function on the cell surface (Hogg, 2003
). The idea that disulfide bonds can act as dynamical redox switches, specifically operated by secreted redox catalysts, represents a novel concept in signal transduction (Jordan and Gibbins, 2006
). However, technical difficulties in detecting and analyzing individual disulfide rearrangements on the cell surface have made progress slow.
Trx1 is recognized as one of the most important regulators of cellular and organismal redox homeostasis (Gromer et al, 2004
). In particular, intracellular Trx1 counteracts oxidative stress, promotes cell growth and inhibits apoptosis. Under conditions of oxidative stress, Trx1 is released by cells and accumulates at sites of inflammation (Nordberg and Arner, 2001
). Numerous studies have reported that secretory Trx1 influences effector functions and proliferation of lymphocytes (Nakamura et al, 1997
). However, proteins and pathways coupling extracellular Trx1 redox activity to defined cellular responses have remained unknown. In this study, we addressed the question regarding which lymphocyte surface receptors are targeted and regulated by the redox activity of extracellular Trx1.
For this purpose, we made use of a mechanism-based kinetic trapping approach to capture mixed disulfide intermediates formed between exogenous Trx1 and its target proteins on the cell surface of living cells. Activity-based techniques offer the opportunity to identify interactions too short-lived to be detectable by conventional methods. To our knowledge, this is the first reported application of kinetic trapping to identify novel target proteins of mammalian Trx1 and the first application of this technique to the surface of intact cells. We demonstrate that Trx1 interacts with intra- and extracellular target proteins in a highly selective manner, guided by specific protein–protein recognition rather than random encounters with disulfide bonds.
Applying the approach to the surface of cell lines representative of the lymphoid lineage, we observed that Trx1 basically targets a single cell surface protein, subsequently identified as TNF receptor superfamily member 8, also known as CD30. The pronounced preference of Trx1 for one particular target protein might seem surprising, but could be due to the fact that we assessed Trx1 reactivity of proteins as they are embedded in their natural microenvironment, namely the intact surface of the active plasma membrane of living cells. It is conceivable that protein disulfide exchange interactions are limited and controlled by their native context and location.
To scrutinize the specificity of the observed interaction, we asked if the preference of Trx1 for CD30 might be caused by an unusual density of disulfide bonds within CD30 and/or exceptional cell surface expression levels. Although the CD30 ectodomain harbors a significant number of predicted disulfide bridges within CRDs, it does not appear to be unusual in terms of disulfide bond composition/density when compared to other members of the superfamily. When tested experimentally, Trx1 failed to interact with other CRD-containing proteins, including the EGFR featuring a total of 25 ectodomain disulfide bonds. The preference for CD30 could not be explained by exceptional surface expression levels either. While Hodgkin's disease cell lines typically express high levels of CD30, other cell lines including LCL-721.220 or CCRF-CEM show at least 20- to 100-fold lower expression as determined by flow cytometry, yet the same selective targeting was observed. Conversely, ectopic overexpression of several related TNFR superfamily members in HeLa cells did not lead to their interaction with Trx1, yet CD30 strongly interacted on the same cells under the same conditions. Consistent with these findings, recent experiments demonstrate that Trx1 targets a particular site within the CD30 ectodomain (Y Balmer and TP Dick, unpublished data).
To facilitate identification of low-abundance cell surface proteins, in vitro trapping experiments were typically performed using Trx1 concentrations of 1–3 μM. However, when disulfide exchange was subsequently tested at lower concentrations, Trx1(CSAAA) concentrations in the low nanomolar range (4–40 nM) were found to give rise to the formation of proportional amounts of Trx1-CD30 mixed disulfide intermediates (), thus demonstrating that the observed interaction is compatible with the expected physiological concentration range of secretory Trx1 (see below).
Wild-type Trx1 is known to act as a multiple-turnover catalyst if a suitable reducing system and electron source is provided for its regeneration. In agreement with these considerations, we observed that sustained reduction of CD30 in cell culture requires a Trx1 regenerating system. Using flow cytometry to monitor conformational changes in the CD30 ectodomain, CD30 was found to respond to Trx1 concentrations in the nanomolar range, starting at around 100 nM (, lower panel). However, the minimal Trx1 concentration required for sustained CD30 reduction might be substantially lower in specific environments, which are efficient in delivering reducing equivalents and preventing oxidative inactivation of Trx1.
At present, it is not clear how extracellular Trx1 is regenerated in vivo
. Despite the overall oxidizing character of the extracellular compartment, reductive processes are known to take place on the cell surface. On the one hand, there is long-standing evidence for the existence of transplasma membrane redox systems delivering electrons to the cell surface (Crane et al, 1985
). On the other hand, Trx1 may be regenerated by co-secreted reductants, as Trx1 secretion in DC-T co-culture is accompanied by the release of reduced cysteine and the creation of a reducing microenvironment between interacting cells (Angelini et al, 2002
). In addition, TrxR was found to be secreted by activated monocytes and might be part of an extracellular Trx1 reducing system (Soderberg et al, 2000
The concentration of Trx1 in human plasma is in the low nanomolar range (1–5 nM), and is found to be elevated several-fold under inflammatory conditions (Yoshida et al, 1999
). However, plasma Trx1 is oxidized and appears to represent systemic dilution of Trx1 previously released within tissues. Accordingly, local tissue concentrations of secretory Trx1, for example, within activated lymph nodes, are expected to be markedly higher than in plasma. Overall Trx1 concentrations in mammalian tissues can be as high as 20 μM (Gromer et al, 2004
). Certain Trx1-secreting cell types, including macrophages and dendritic cells, distinctly upregulate expression of Trx1 upon activation (Angelini et al, 2002
). In vitro
studies of Trx1 secretion suggest that a substantial fraction of intracellular Trx1 can be released within a few hours (Rubartelli et al, 1992
). Although direct measurements of extracellular Trx1 within tissues are not available, physiologically relevant extracellular Trx1 concentrations may well reach into the upper nanomolar, if not lower micromolar range.
We found that Trx1-mediated disulfide reduction changes the conformation and functional properties of the CD30 ectodomain. In the reduced state, CD30 lost its ability to interact with its cognate ligand CD30L or agonistic antibodies. The presence of catalytically active Trx1 impeded CD30-dependent signaling in the YT lymphoma cell line, as demonstrated by its effect on CD25 and FasL expression, as well as its influence on cytotoxicity against Fas-expressing target cells.
The physiological role of the CD30-CD30L system has remained unclear. In vitro
studies focusing on CD30+
lymphoid malignancies showed that triggering of CD30 signaling can induce either proliferation, activation, growth arrest or apoptosis, depending on cell type and stimulatory conditions (Schneider and Hubinger, 2002
). In vivo
, cell surface expression of CD30 appears to be tightly regulated and restricted to B and T lymphocytes undergoing activation in lymphoid tissues. It has been proposed that CD30 provides proliferation and/or survival signals during lymphocyte responses (Croft, 2003
Under inflammatory conditions, CD30 expression is markedly induced. In vivo
activation of CD30 can be monitored by the release of sCD30, shed from the plasma membrane upon CD30L binding (Hansen et al, 2000
). Similar to serum Trx1, serum sCD30 is increased in infection, autoimmunity and allergy, for example systemic lupus erythematosus, rheumatoid arthritis and atopic dermatitis (Horie and Watanabe, 1998
). Both Trx1 and CD30 appear to play a role in the regulation of the antiviral inflammatory response. Both Trx1 secretion and CD30 expression have been associated with virally transformed lymphocytes. Elevated levels of sCD30 occur during viral infection. Likewise, viral infection leads to elevated Trx1 plasma levels and several studies indicate that secreted Trx1 modulates the antiviral inflammatory process (Nakamura et al, 2001b
In this study, we have identified an enzyme–substrate relationship between Trx1 and CD30, a receptor of activated lymphocytes involved in the regulation of inflammation. As lymphocytes migrate between different microenvironments, it is conceivable that Trx1 catalyzes disulfide exchange dynamically, activating or inactivating the CD30 pathway in response to the redox environment. The interaction between Trx1 and CD30 might represent a regulatory link between oxidative stress and lymphocyte function. Understanding of this relationship in vivo awaits the generation of suitable experimental tools.