The family of collapsin response mediator proteins (CRMPs) is integral phosphoproteins that are developmentally regulated within the brain and share substantial homology with the dihydropyrimidinase liver enzyme, that acts as a hydrolase of dihydrouracil. The CRMPs have been defined as CRMP1-4, which share approximately 70–75% of amino acid sequences and that can form oligomers with each other, regulating numerous cytosolic events within neural or other cell lineages but lack the enzymatic activity of dihydropyrimidinase. The newest defined member of this family, CRMP-5 has the least sequence similarity of these phosphoproteins sharing only approximately 50% similarity but has been highlighted as a potential marker in various cancers, which consist of neurological syndromes and may also take center stage during neurodevelopment (Taghian et al. 2012). In this issue of the Journal of Neurochemistry, Ponnusamy and Lohkamp now report for the first time, the crystal structure of the full-length and a truncated form of CRMP-5 mapping out how this protein interacts with itself as a homotetramer but importantly, how it can compete and heteromerize with greater affinity to CRMP-2 and to a lesser extent CRMP-1. Despite the structural similarities defined in this study for CRMP-5 when compared with CRMP-1 and -2, the described differences at interaction interfaces, from positively charged to uncharged residues, demonstrate the propensity for hetero-oligomerization with CRMP-2 rather than CRMP-1. This novel finding may suggest an important physiological role for CRMP-5 regulating the other developmentally regulated CRMP family members, with implications for drug targeting of the binding motifs during development and disease (Ponnusamy and Lohkamp 2013).
Through the crystallization of CRMP-5, Ponnusamy and Lohkamp show that the amino acid residues 483–492 can form an extended loop from the protomer to the other monomer with a salt bridge formed between the guanidine group of R489 as well as at the carboxyl group E214, providing structural integrity for the oligomer. This dimer interface was shown to consist of an area, which included 10 hydrogen bonds and salt bridges, termed an ‘arm-lever’ co-existing with an alternate structure termed ‘arm-arm’ consisting of an area that includes 16–18 hydrogen bonds and salt bridges (Fig. 1). Both structures were shown to potentiate CRMP-5 homotetramers and correspond favorably with recent reports on the crystal structures of CRMP-1 and -2 (Ponnusamy and Lohkamp 2013). However, the authors were able to show that CRMP-5 preferentially interacts with CRMP-2 than CRMP-1 when forming heterotetramers(Fig. 1). It is reported that at the ‘arm-lever’ interface, the side chains of N237 and K265 both form hydrogen bonds with E223. Since CRMP-1 and -2 have a proline and threonine at the K265 position, there would appear to be an inability to form hydrogen bonds at this interface. However, CRMP-2 maintains the N237 at the α-helix 7, whereas CRMP-1 displays a glycine instead, suggesting that at least one of the interfaces can form hydrogen bonds between CRMP-5 and -2 (Fig. 1). Furthermore in the CRMP-5 arm-arm interface, 3 hydrogen bonds can be formed and this is conserved in all other CRMP family members except for CRMP-1 where K473 is replaced by glutamine. For these reasons, it would seem that there exists weaker affinity at the arm-arm and arm-lever interfaces for the formation of CRMP-5 and -1 oligomers. These data now implicate CRMP-5 as a competitive binding partner for CRMP-2 during hetero-oligomerization (Ponnusamy and Lohkamp 2013).
The authors were able to consolidate the effect of the divalent cations Ca2+ and Mg2+, on CRMP-2 homotetramer formation (Ponnusamy and Lohkamp 2013). The data illustrating tetramer stability through cation binding were previously reported by Majava et al. (2008) who could demonstrate the prevention of β-aggregate formation upon heating in the presence of Ca2+ and Mg2+. In this study, the binding sites of these cations were shown to be overlapping and on the tetramer surface, away from dimer interfaces. They showed that in CRMP-5, there is a replacement of residue 349 on the main chain carbonyl oxygen and side chains of E353 or Q81 of CRMP-2, integral for Ca2+ binding, with H346 and M71, respectively, thereby preventing size and charge interaction sites for the cation. In CRMP-2, the binding of cations at these sites is important as they are close to the side chain of Q245, which stabilizes the CRMP-2 monomer assisting tetramerization (Fig. 1). These interactions are not found for either CRMP-5 or CRMP-1 implicating these family members in heteromeric interactions with CRMP-2. Furthermore, the relevance of CRMP-2 homotetramer formation is important in developmental neurobiology with Ca2+ being an integral cation that regulates axonogenesis and guidance (Fig. 1). Since the importance of tetrameric CRMP-2 in growth-related cargo vesicle transport has been postulated for axon growth and plasticity (Kimura et al. 2005; Tsuboi et al. 2005; Arimura et al. 2009), such structural data may be important in defining these mechanisms of neuronal development. This may also mean that destabilizing such tetrameric structures may be involved in the failure of axon growth (Fig 1).
One critically unresolved aspect of all structural studies surrounding the CRMPs is the integrity of the C-terminal region. In this study once again, the C-terminus (last 84 amino acids) of CRMP-5 could not be visualized but is suggested to be unstructured. It is however this domain which has had substantial interest with regard to the physiological function of all CRMPs. The C-terminal domain contains most, if not all, of the phosphorylation sites regulated by various kinases and also consist of putative calpain cleavage sites. Both of these post-translational modifications for all of the CRMPs, including CRMP-5, can have profound effects on the ability to bind tubulin heterodimers as well as various other protein binding partners. These events have been well documented to contribute to the abrogation of neurite outgrowth and even neurotoxicity (Taghian et al. 2012). Despite the uncertainty regarding the C-terminus it does not seem to alter the tetrameric assembly of CRMP-5 since it is likely that it extends away from its own protomer, ineffective in causing structural changes to the core of the protein.
Despite the intense research interest in the CRMP family of phosphoproteins for over a decade, most of the focus has been directed toward the phosphorylation modifications and novel binding partners of these molecules defining outcomes in developmental neurobiology (Taghian et al. 2012). Previous elucidation of the crystal structure in the most profoundly characterized member, CRMP-2, has identified a total buried solvent-accessible surface area for the homotetramer as being over 9400 Å2, with physiological significance. Furthermore, the generation of heterotetramers between CRMP-1 and -2 have revealed conserved residues with 3 sequence differences (CRMP-1: Leu266 and Tyr316 vs. CRMP-2: Gln266 and Phe316) at interface 1, thereby demonstrating interaction affinity (Stenmark et al. 2007).
The crystal structure of the newest member of this family, CRMP-5, has now been elucidated by Ponnusamy and Lohkamp who uncover significant complexity in the family of CRMP proteins, demonstrating homo- and hetero-oligomerization propensities with direct impact on the physiological outcomes of these assemblies. Their findings may now direct research into the nature of homo- and/or hetero-oligomerization at a fundamental level during neural cell development, or during the development and integration of the nervous system with possible implications for neurodegenerative diseases.