We present a thorough analysis of nuclear export signals and a prediction server, which we have made publicly available. The machine learning prediction method is a significant improvement over the generally used consensus patterns. Nuclear export signals (NESs) are extremely important regulators of the subcellular location of proteins. This regulation has an impact on transcription and other nuclear processes, which are fundamental to the viability of the cell. NESs are studied in relation to cancer, the cell cycle, cell differentiation and other important aspects of molecular biology. Our conclusion from this analysis is that the most important properties of NESs are accessibility and flexibility allowing relevant proteins to interact with the signal. Furthermore, we show that not only the known hydrophobic residues are important in defining a nuclear export signals. We employ both neural networks and hidden Markov models in the prediction algorithm and verify the method on the most recently discovered NESs. The NES predictor (NetNES) is made available for general use at http://www.cbs.dtu.dk/.

Protein export from the nucleus is often mediated by a Leucine-rich Nuclear Export Signal (NES). NESbase is a database of experimentally validated Leucine-rich NESs curated from literature. These signals are not annotated in databases such as SWISS-PROT, PIR or PROSITE. Each NESbase entry contains information of whether NES was shown to be necessary and/or sufficient for export, and whether the export was shown to be mediated by the export receptor CRM1. The compiled information was used to make a sequence logo of the Leucine-rich NESs, displaying the conservation of amino acids within a window of 25 residues. Surprisingly, only 36% of the sequences used for the logo fit the widely accepted NES consensus L-x(2,3)-[LIVFM]-x(2,3)-L-x-[LI]. The database is available online at http://www.cbs.dtu.dk/databases/NESbase/.

Numerous, highly conserved RING-H2 domains are found in the model plant Arabidopsis thaliana (thale cress). To characterize potential RING-H2 protein interactions, the small RING-H2 protein RHA2a was used as bait in a yeast two-hybrid screen. RHA2a interacted with one of the plant-specific NAC [NAM ('no apical meristem'), ATAF1/2, CUC2 ('cup-shaped cotyledons 2')] transcription factors, here named ANAC (abscisic acid-responsive NAC). The core RING-H2 domain was sufficient for the interaction. The ability of 11 structurally diverse RING-H2 domains to interact with ANAC was then examined. Robust interaction was detected for three of the domains, suggesting multi-specificity for the interaction. The domains that interacted with ANAC contain a glutamic acid residue in a position corresponding to a proline in many RING-H2 domains. Conversion of this glutamic acid residue into proline in RHA2a decreased its ability to interact with ANAC, most likely by changing the interaction surface. This suggested that a short, divergent region in RING-H2 domains modulate interaction specificity. ANAC contains a degenerate bipartite nuclear localization signal (NLS), while RHG1a, also identified as an ANAC interaction partner, contains a basic NLS. Both signals localized beta-glucuronidase reporter fusions to the nucleus. N-terminally truncated RHA2a also directed nuclear localization, apparently dependent on basic amino acids in the RING-H2 domain. Nuclear co-localization of the RING-H2 proteins and ANAC may enable their interaction in vivo to regulate the activity of the ANAC transcription factor.

Classic FYVE zinc-finger domains recognize the phosphoinositide signal PtdIns3P and share the basic (R/K)(1)(R/K)HHCR(6) (single-letter amino acid codes) consensus sequence. This domain is present in predicted PtdIns3P 5-kinases and lipases from Arabidopsis thaliana. Other Arabidopsis proteins, named PRAF, consist of a pleckstrin homology (PH) domain, a regulator of chromosome condensation (RCC1) guanine nucleotide exchange factor repeat domain, and a variant FYVE domain containing an Asn residue and a Tyr residue at positions corresponding to the PtdIns3P-interacting His(4) and Arg(6) of the basic motif. Dot-blot and liposome-binding assays were used in vitro to examine the phospholipid-binding ability of isolated PRAF domains. Whereas the PH domain preferentially bound PtdIns(4,5)P(2), the variant FYVE domain showed a weaker charge-dependent binding of phosphoinositides. In contrast, specificity for PtdIns3P was obtained by mutagenic conversion of the variant into a classic FYVE domain (Asn(4),Tyr(6)-->His(4),Arg(6)). Separate substitutions of the variant residues were not sufficient to impose preferential binding of PtdIns3P, suggesting a co-operative effect of these residues in binding. A biochemical function for PRAF was indicated by its ability to catalyse guanine nucleotide exchange on some of the small GTPases of the Rab family, permitting a discussion of the biological roles of plant FYVE proteins and their regulation by phosphoinositides.