3.1. Htt Structure and Function
In HD patients, the polyQ region located after the first seventeen N-terminal Htt residues is expanded beyond a threshold of 36 glutamine residues (mHtt)
[34].
A distinctive feature of HD is the progressive death of striatal projection neurons (SPNs), which are GABAergic output neurons representing >90% of striatal cells. SPNs are divided into two groups, depending on whether they belong to the direct (DP) or indirect (IP) pathway (DP-SPNs and IP-SPNs, respectively). Both SPN sub-types receive extensive glutamatergic inputs from cortex and thalamus, and dopaminergic inputs from the ventral tegmental area and substantia nigra pars compacta
[34].
Within neurons, mHtt molecules form toxic aggregates with a rate proportional to the length of the polyQ expansion. mHtt co-aggregates with a number of different proteins, which decreases their concentration in the cell. Several studies demonstrate that polyQ expansion in mHtt results in a toxic gain-of-function phenotype. Other studies, which include gene knockouts and knockdowns, demonstrate that polyQ expansion in mHtt can also have loss-of-function effects. The role of wild-type Htt should therefore be taken into account in the development of a potential therapy, as well as that of mHtt
[35].
The structure of Htt in complex with HAP40 has been recently solved by cryo-EM at 4 Å resolution
[17]. Conversely, unbound Htt is very difficult to study either by X-ray crystallography or cryo-EM because it has a large size (it contains 3144 residues and weights 348 kDa), is very flexible, and tends to form aggregates. However, about 28% of Htt residues are not visible in the structure, including the first 91 residues, which comprise the 17 glutamine residue-containing polyQ expansion. In agreement with computational predictions, all secondary structure elements belonging to either Htt or HAP40 resolved in the model are α-helices, most of which (72%) are arranged in HEAT or other tandem repeats.
The cryo-EM structure revealed that Htt is formed by three domains: N-terminal domain and C-terminal domain, both of which contain multiple HEAT repeats, joined to a bridge domain (Figure 4).
Figure 4. X-ray structure of the complex between HAP40 and Htt (PDB code: 6EZ8). HAP40 is colored green and the C-Heat, Bridge, and N-Heat domains of Htt are colored orange, yellow, and cyan, respectively.
The N-terminal domain (N-HEAT; residues 91–1684) forms a typical α-solenoid, comprising 21 HEAT repeats arranged as a one-and-a-half turn, right-handed superhelix. The N-HEAT was predicted to contain two membrane binding regions: the 1–17 N-terminal tail (not visible in the structure), which was predicted to form an amphipathic helix
[36], and a larger region (comprising a.a. 168–366), containing a functionally important palmitoylation site at Cys208
[37]. This region (N-HEAT repeats 2–4) is placed on the N-HEAT convex surface and is positively charged.
The C-HEAT (a.a. 2092–3098) comprises 12 HEAT repeats that form an elliptical ring of ~80 × 30 Å.
The N-HEAT and C-HEAT are joined by the bridge domain. This contains six tandem α-helical repeats, of which repeats three, four and six are Armadillo-like. The repeat region is flanked by five non-repeat helices and a flexible C-terminus (a.a. 2062–2092), which is unresolved.
The lack of interactions between the N-HEAT and C-HEAT domains explains the flexibility of the Htt structure in the absence of interactors. HAP40 binds to a large cleft between these two domains. Within the complex, Htt and HAP40 share large interfaces that comprise mostly hydrophobic interactions. Interestingly, bioinformatic analyses has provided evidence of the evolutionary conservation of these interfaces in all Htt-encoding animal species
[38].
Recently, the structure of wild-type Htt in complex with HAP40 has been compared with other mHtt–HAP40 complexes differing in the length of the polyQ repeat, namely, 46QHtt–HAP40 (46 being a typical polyQ length in HD patients) and 128QHtt–HAP40 (128 being an extremely high polyQ length). Quite surprisingly, both crosslinking mass spectrometry and cryo-EM experiments revealed no major structural differences among the different complexes, indicating that the polyQ insertion does not alter the Htt fold
[39].
Htt function has not yet been fully elucidated, but Htt has been proposed to be involved in cellular processes such as mitotic spindle orientation, autophagy, and vesicle transport
[40][41][42][43]. Htt has been shown to be essential in both embryonal development and adult life. In mice, the Htt knockdown mutant dies at about day 8.5 of gestation
[44]. Additionally, Htt deletion in the mouse central nervous system leads to a phenotype similar to that of HD, i.e., cellular stress, neuroinflammation, aberrant synaptic connectivity, and neuronal death
[45][46][47].
Htt is crucial for energetic metabolism not only in the brain
[48] but also in peripheral tissues. In Htt-null cardiomyocytes, both intracellular ATP and total purine concentration in the cellular medium were reduced. This indicates that, in the heart, Htt plays an important role in both cellular energy balance and nucleotide metabolism
[49].
Recent studies carried out using mouse embryonic stem cell demonstrated that Htt is necessary for mitochondrial structure and function from the earliest stages of embryogenesis, providing a molecular explanation for the early embryonic lethality of Htt knockdown
[50].
SPN death in HD has mostly been ascribed to toxic ‘gain-of-function’ by mHtt. However, evidence of Htt ‘loss-of-function’ contribution is also available. Indeed, wild-type Htt has been shown to be neuroprotective and, therefore, able to shield neurons against mHtt toxicity
[51]. Additionally, Htt deletion in IP-SPN and DP-SPN leads to a phenotype that resembles the key features of HD, supporting the hypothesis that Htt loss-of-function contributes to SPN pathology in HD
[52]. Lowering of wild-type Htt expression has also been shown to affect both health and function of primary monocyte-derived macrophages from healthy human subjects, likely by different mechanisms with respect to those associated with mHtt
[53].
3.2. Htt Ubiquitination and SUMOylation
Htt is a protein with high ubiquitination potential. It contains 124 lysine residues, many of which are placed on the protein surface, as shown by cryo-EM structural analysis, and may in principle be ubiquitinated or linked to other Ub-like (Ubl) proteins, such as SUMO (Small Ub-like Modifier).
Ubiquitination and/or SUMOylation has been demonstrated for the 30 Htt lysine residues reported in
Table 1. Ubiquitination and SUMOylation processes have been shown to compete for lysine residues K6, K9, and K15, all of which are placed in the N-terminal tail that is not visible in the solved cryo-EM structure
[54], and they may compete for other lysine residues as well.
Table 1. List of ubiquitinated and SUMOylated residues in human Htt, according to PhosphoSitePlus (phosphosite.org).
Residue (Human Htt) |
Demonstrated Modification |
Reference |
Lys 6 |
SUMOylation |
[54] |
Lys 9 |
SUMOylation |
[54] |
Lys 253 |
Ubiquitination |
[55] |
Lys 335 |
Ubiquitination |
[56] |
Lys 442 |
Ubiquitination |
[57] |
Lys 662 |
Ubiquitination |
[56] |
Lys 667 |
Ubiquitination |
[55] |
Lys 698 |
Ubiquitination |
[55] |
Lys 813 |
Ubiquitination |
[55] |
Lys 902 |
Ubiquitination |
[55] |
Lys 937 |
Ubiquitination |
[55] |
Lys 941 |
Ubiquitination |
[58] |
Lys 1121 |
Ubiquitination |
[56] |
Lys 1223 |
Ubiquitination |
[56] |
Lys 1244 |
Ubiquitination |
[56] |
Lys 1262 |
Ubiquitination |
[59] |
Lys 1337 |
Ubiquitination |
[55] |
Lys 1402 |
Ubiquitination |
[55][56][58][59][60] |
Lys 1410 |
Ubiquitination |
[55][56][58] |
Lys 1415 |
Ubiquitination |
[56] |
Lys 1431 |
Ubiquitination |
[56][58][61] |
Lys 1885 |
Ubiquitination |
[55] |
Lys 2417 |
Ubiquitination |
[59] |
Lys 2423 |
Ubiquitination |
[56] |
Lys 2443 |
Ubiquitination |
[56] |
Lys 2537 |
Ubiquitination |
[55] |
Lys 2564 |
Ubiquitination |
[55] |
Lys 2757 |
Ubiquitination |
[55] |
Lys 2901 |
Ubiquitination |
[55] |
Lys 2967 |
Ubiquitination |
[55] |
Ubiquitination has been linked to reduced mHtt toxicity, most likely due to increased mHtt clearance by the proteasome
[62]. Conversely, SUMOylation has been shown to stabilize mHtt, reduce mHtt aggregation, enhance transcriptional dysregulation by mHtt, and increase mHtt toxicity in a Drosophila model
[54]. SUMOylation contribution to mHtt toxicity may be mediated by mHtt targeting of the nucleus and the hampering of ubiquitination and subsequent degradation. While Htt ubiquitination is mediated by different E3 ligases, mHtt SUMOylation is mediated by only one protein, i.e., RHES (Ras homolog enriched in striatum)
[63]. RHES has higher affinity for mHtt than wild-type Htt, and its selective expression in the striatum strongly suggests that this protein contributes to the HD pathology
[64]. For these reasons, RHES has been considered to be an attractive target for HD therapy
[65].
As described above, the presence of seven lysine residues allows Ub to transmit different signals on target proteins. Therefore, Htt polyubiquitination can take place with different mechanisms, which have not yet been completely understood. Two among the 40 E2 enzymes present in the human genome have been demonstrated to interact with Htt: UBE2W, which is known to target the N-termini of disordered proteins such as Tau, RBPB8, and ATXN3, as well as mHtt
[66]; and UBE2K, also known as Htt-interacting protein (Hip2) or E2-25k, which is a Ub-conjugating enzyme that directly interacts with Htt and may mediate its ubiquitination
[67].