Nucleophosmin mutations in acute myeloid leukemia: a tale of protein unfolding and mislocalization

Abstract

Nucleophosmin (NPM1) is an abundant, ubiquitously expressed protein mainly localized at nucleoli but continuously shuttling between nucleus and cytoplasm. NPM1 plays a role in several cellular functions, including ribosome biogenesis and export, centrosome duplication, chromatin remodeling, DNA repair, and response to stress stimuli. Much of the interest in this protein arises from its relevance in human malignancies. NPM1 is frequently overexpressed in solid tumors and is the target of several chromosomal translocations in hematologic neoplasms. Notably, NPM1 has been characterized as the most frequently mutated gene in acute myeloid leukemia (AML). Mutations alter the C-terminal DNA-binding domain of the protein and result in its aberrant nuclear export and stable cytosolic localization. In this review, we focus on the leukemia-associated NPM1 C-terminal domain and describe its structure, function, and the effect exerted by leukemic mutations. Finally, we discuss the possibility to target NPM1 for the treatment of cancer and, in particular, of AML patients with mutated NPM1 gene.

Figure 1

Schematic representation of NPM1 functional domains. Two leucine-rich nuclear export signals (NES), recognized by the exportin1-Crm1 transport system, are present in the N-terminal chaperone domain. A bipartite nuclear localization signal (NLS) is found at the central unstructured domain. Finally, a nucleolar localization signal (NoLS) is found at the C-terminal domain. The black bar indicates the region endowed with ribonuclease activity. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.] [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 2

NPM1 C-terminal domain structure. (A) The terminal three-helix bundle is characterized by a small hydrophobic core mainly constituted by residues Phe268, Phe276, Trp288, and Trp290. The two tryptophan residues (or Trp290 alone) are mutated in the leukemia-associated variants. (B) The three-helix bundle here is shown in a different orientation to highlight the positive charges that are distributed throughout the surface of the domain. Residues Lys250, Lys257, Lys267, Asn270, Asn274, and Cys275 are located at the interface with G-quadruplex DNA. (C) Sequence alignment of the DNA-binding domain. Residues implicated in the formation of the hydrophobic core (in magenta) and in contacting G-quadruplex DNA (highlighted in blue, green, and yellow boxes for lysines, asparagines, and cysteine, respectively) are strictly conserved. An unstructured tail that precedes the three-helix bundle is necessary to confer high affinity for DNA and is also enriched in lysine residues (in cyan). Several residues in the tail, and at the interface with DNA, are subjected to posttranslational modifications: phosphorylation, acetylation, and sumoylation are highlighted with blue, red, and green stars, respectively. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.] [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 3

Mutations at the C-terminal domain are associated with acute myeloid leukemia. Here, the most common mutations are shown. They consist in the insertion or duplication of small base stretches (underlined). For clarity, one every two codons in the terminal sequence of the gene is highlighted in bold and the corresponding amino acids are indicated in gray. The resulting amino acidic sequences are also shown in the right column. The altered reading frame leads to the expression of mutated proteins that have a different C-terminal sequence and are longer by four residues (in italics). Both Trp288 and Trp290 (or Trp290 only in mutants E and F) are replaced and a new nuclear export signal (residues in gray boxes in the right column) appears in all cases.

Figure 4

A G-quadruplex sequence from the c-MYC promoter is shown in complex with the NPM1 DNA-binding domain (model derived from NMR data). The G-quadruplex backbone is accommodated into a small positive groove formed by helices H1 and H2. Several interactions between lysine and asparagine residues and the DNA backbone phosphates stabilize the complex. The N-terminal tail is still unstructured in the complex even though it proved necessary for high-affinity binding. Lysine residues in the tail are highlighted. Mutation of Lys229 and Lys230 to alanine dramatically reduces the binding affinity. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.] [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 5

Proposed strategies for targeting NPM1 in AML. (A) As wild-type NPM1 oligomerizes with mutated NPM1 and is thus translocated to the cytosol, only a small fraction of the protein is retained in nucleoli. A drug able to dissociate residual wild-type NPM1 from nucleoli might cause nucleolar stress and disruption, ultimately leading to apoptosis in leukemic blasts. (B) An opposite strategy would be that of developing a drug acting as a pharmacological chaperone. Such drug should trigger mutated NPM1 refolding upon binding and shift, through mass effect, the folded–unfolded equilibrium toward the folded species. This would in turn facilitate the relocalization of both native and mutated NPM1 in nucleoli, thus counteracting the antiapoptotic effects played by NPM1c+. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.] [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]