HD-tuin proteins, the ugly ducklings of histone deacetylases

In Arabidopsis thaliana, the HD-tuin (HDT) type of histone deacetylase (HDAC) family is formed by four peculiar genes. On the one hand, these could be considered greener than other members of the family, since they are only found in plants and could have evolved to perform plant-specific functions (Hollender and Liu, 2008). On the other hand, much like the ‘ugly duckling’ in the fairy tale, they do not fit in with their HDAC siblings, having no sequence similarity with other HDACs. Instead, they resemble the FK506-binding protein (FKBP) nucleoplasmin, family of histone chaperones that store nucleosomal histones or exchange them during chromatin remodeling and other processes like gene expression. In the new work, Ruchir C. Bobde and colleagues (Bobde et al. 2022) have structurally characterized the N-terminal domain (NTD) of the four HDTs and have reclassified them. The authors showed that the ugly HDTs turned out to be swans, or in this case, HDT nucleoplasmins.

The authors crystallized the NTD of HDT1–4 proteins and found all HDTs form a pentamer, with a hole in the middle (Figure). Each monomer is composed of eight antiparallel beta strands, and the monomer–monomer interaction is stabilized by hydrophobic interactions, hydrogen bonds, and van der Waals contacts. The structure giving a better match on the PDBeFold server was the nucleoplasmin fold. This confirmed the predictions of the phylogenetic analysis, where HDTs cluster together with FKBP nucleoplasmins, and expectations from previous studies (Aravind and Koonin, 1998; Edlich-Muth et al., 2015).

Figure.

Pentameric structure with a doughnut shape of the NTD of nucleoplasmins HDT1–4. The monomers are composed of beta sheets (Adapted from Bobde et al. (2022), Figure 2A).

To test if the properties of HDTs resemble those of nucleoplasmins, the authors measured their thermal stability and resistance to denaturation and proteases. As nucleoplasmins, HDTs were highly thermostable and remained as a pentamer even after withstanding high temperatures. They also coped well with high concentrations of urea and salts, proving highly resilient to denaturing agents. Finally, HDTs also seemed to be quite immune to protease digestion. All these features show that HDTs and nucleoplasmins are similarly stable.

But do these swans have some similarity to ducks? That is, can HDTs deacetylate histones? The four HDTs have triads of amino acids predicted to be key for this function oriented toward the core of the pentamer. However, there was no deacetylation activity in vitro using recombinant HDT NTDs. The triads seem to be catalytically inactive possibly because of the way the amino acids are positioned, slightly out of range. In contrast, HDTs have been reported to have deacetylation activity in vivo in Arabidopsis and other plants, with hdt knockouts having higher histone acetylation levels (Wu et al., 2000; Ding et al., 2012). It is possible that cofactors, other protein partners, or even posttranslational modifications are missing when the deacetylation activity is analyzed in vitro. Therefore, it is still unclear if HDTs are unable to deacetylate histones, or they only do it under certain conditions.

The next step was testing through analytical gel filtration chromatography if HDTs could interact with histones, histone dimers, or nucleosomes. Although the NTD alone could not interact with histones, an extended version of HDT2 including the acidic tract interacted with the H2A/H2B dimer and with the H3/H4 tetramer in a 1:1 ratio. These results highlight the importance of the acidic tract for the interaction with histones and reinforce the function of HDTs as nucleoplasmins. Unlike the FKBP nucleoplasmin, FKBP53, HDT2 did not interact with preformed nucleosomes, which suggests a specific function for the HDT proteins.

The study by Bobde and colleagues confirms that the plant-specific HDT proteins are structurally nucleoplasmins and able to chaperone histone H2A/H2B dimers and H3/H4 tetramers. Further studies are needed to evaluate the implications of this function in vivo and whether it could be combined with HDAC activities in planta.