Abstract

Ubiquitination of lysine (K) residues in the cytoplasmic domain of cytokine receptors plays a major role in intracellular receptor routing and control of signal duration. However, which ubiquitin (E3) ligases and whether specific lysines, eg. present in conserved motifs, are involved in ubiquitin-mediated routing is still largely unknown. We recently showed that SOCS3, which forms an Elongin/Cullin-based E3 ligase (ECSSOCS3), is involved in ligand-induced ubiquitination, lysosomal routing and degradation of the G-CSFR (

Irandoust et al,
EMBO J
.
2007
,
26
:
1782
–93
). Strikingly, ECSSOCS3-mediated mechanisms largely depended on ubiquitination of a single juxtamembrane lysine at position 632 (K632), even though 4 additional conserved lysines are present in the cytoplasmic domain of the G-CSFR. This juxtamembrane lysine located 5 amino acids upstream of the box1 region is conserved among several other cytokine receptors. We sought to unravel the configuration required for ubiquitination of K632 and for its function in attenuation of G-CSF signaling. First, we tested whether K632 is part of a domain that can function as an isolated motif. To this end, we fused the K632-containing domain (D) encompassing the juxtamembrane and box1 regions to the COOH-terminus of a lysine-less G-CSFR (K5R-D). Despite efficient ubiquitination of the lysine within the reallocated domain, K5R-D, similar to K5R, was severely hampered in lysosomal routing. This resulted in prolonged STAT5 activation and in G-CSF-induced hyperproliferation of myeloid 32D cells indicating that the juxtamembrane domain of the G-CSFR does not function as an isolated domain that can be shifted to a different cytoplasmic location. To delineate how imperative the positioning of K632 is for its function, we inserted 5 alanines immediately upstream or downstream of K632. However, these insertions did not affect G-CSFR signal duration suggesting that there is no stringent proximity of K632 to the cell membrane and no strict positioning relative to the SOCS3 recuitment site (Y729) to direct downregulation of G-CSFR signaling. To further study this flexibility in positioning we ‘walked’ lysines through the juxtamembrane domain and the adjacent box 1-region by mutating amino acids at indicated positions to a lysine in an otherwise lysine-less G-CSFR. Mutants 628K and 630K were hardly ubiquitinated and, as predicted, were hampered in lysosomal routing. In contrast, lysines on position 631, 633, 634, 638 and 644 were ubiquitinated comparably to K632. Interestingly, while mutants 631K, 633K and 634K showed characteristics indistinguishable from 632K, mutants 638K and 644K displayed slightly decreased signal attenuation suggesting a gradual decline of functionality of lysines moved to box-1. Finally, moving the lysine downstream to position 672, which corresponds to the second lysine in the wt G-CSFR, resulted in complete loss of functionality. In conclusion, these results indicate that the positioning of a lysine within the juxtamembrane domain is crucial for ECSSOCS3-mediated signal downregulation and lysosomal routing of the G-CSFR, although a limited flexibility is tolerated towards the box1 region. The fact that inactive lysines are still ubiquitinated implies that interaction with effector proteins is involved in lysosomal routing rather than reduced or lost activity of E3 ligases. Given its high conservation among several cytokine receptors we are currently investigating a common role for the juxtamembrane lysine in cytokine receptor routing and signal attenuation.

Author notes

Disclosure: No relevant conflicts of interest to declare.