NAA10 p.(D10G) and NAA10 p.(L11R) Variants Hamper Formation of the NatA N-Terminal Acetyltransferase Complex

The majority of the human proteome is subjected to N-terminal (Nt) acetylation catalysed by N-terminal acetyltransferases (NATs). The NatA complex is composed of two core subunits—the catalytic subunit NAA10 and the ribosomal anchor NAA15. Furthermore, NAA10 may also have catalytic and non-catalytic roles independent of NatA. Several inherited and de novo NAA10 variants have been associated with genetic disease in humans. In this study, we present a functional analysis of two de novo NAA10 variants, c.29A>G p.(D10G) and c.32T>G p.(L11R), previously identified in a male and a female, respectively. Both of these neighbouring amino acids are highly conserved in NAA10. Immunoprecipitation experiments revealed that both variants hamper complex formation with NAA15 and are thus likely to impair NatA-mediated Nt-acetylation in vivo. Despite their common impact on NatA formation, in vitro Nt-acetylation assays showed that the variants had opposing impacts on NAA10 catalytic activity. While NAA10 c.29A>G p.(D10G) exhibits normal intrinsic NatA activity and reduced monomeric NAA10 NAT activity, NAA10 c.32T>G p.(L11R) displays reduced NatA activity and normal NAA10 NAT activity. This study expands the scope of research into the functional consequences of NAA10 variants and underlines the importance of understanding the diverse cellular roles of NAA10 in disease mechanisms.

Here, we report the functional consequences of two de novo NAA10 variants, c.29A>G p.(D10G) and c.32T>G p.(L11R). NAA10 c.29A>G p.(D10G) was previously identified in a boy who died within 6 months of age [40]. The individual had DD, brain malformations, cardiac abnormalities, hearing loss and dysmorphic features. NAA10 c.32T>G p.(L11R) was previously described in a young girl with DD, brain malformations and strabismus [40].

Analysis of NAA10 Sequence and NatA Structure
NAA10 is evolutionarily conserved from yeast to human [1]. A NAA10 multiple sequence alignment shows that the affected amino acids D10 and L11 are highly conserved, suggesting an important role in the function and/or stability of NAA10 ( Figure 1a). The neighbouring residues D10 and L11 are located in the NAA10 α1 helix, which is part of the NAA10-NAA15 hydrophobic binding surface (Figure 1b). DynaMut [44] was used to identify potential interactions mediated by D10 and L11 with nearby amino acids. The negatively charged D10 was predicted to interact with the positively charged R4 in NAA10 β1 through electrostatic and hydrogen bonding. L11 was predicted to form a hydrophobic pocket together with P8, M14, M28, F32 and Y31. The residue R4 interacting with D10 and the hydrophobic residues interacting with L11 are also conserved in many NAA10 orthologues (Figure 1a), supporting a hypothesis that NAA10 p.(D10G) and p.(L11R) variants may disrupt important interactions necessary for proper NAA10 function and/or stability. D10, L11 and the NAA10 residues they interact with were not predicted to directly interact with any NAA15 residues.

Complex Formation and Catalytic Activity
To explore the impact of NAA10-D10G-V5 and NAA10-L11R-V5 on NatA complex formation and catalytic activity, HeLa cells overexpressing either the variants or wildtype NAA10 (NAA10-WT) were subjected to anti-V5 immunoprecipitation (IP), Western blot analysis and Nt-acetylation assays. A striking observation from the Western blot analyses was that less NAA15 co-immunoprecipitated with both variants as compared to NAA10-WT-V5 (Figure 2a,b; Figure S1). These results suggest that both variants impede NatA complex formation, which ultimately is likely to cause reduced NatA-mediated Nt-acetylation in vivo.  [21] bound with ligand IP6 (magenta). The human NatA structure was superimposed with the substrate (SASE, yellow sticks; CoA, pink sticks) from the S. pombe NatA structure (PDB ID: 4KVM) [15]. A zoomed-in view (right) shows D10 (purple) and L11 (orange) and the interactions they form with surrounding amino acids.

Complex Formation and Catalytic Activity
To explore the impact of NAA10-D10G-V5 and NAA10-L11R-V5 on NatA complex formation and catalytic activity, HeLa cells overexpressing either the variants or wildtype NAA10 (NAA10-WT) were subjected to anti-V5 immunoprecipitation (IP), Western blot analysis and Nt-acetylation assays. A striking observation from the Western blot analyses was that less NAA15 co-immunoprecipitated with both variants as compared to NAA10-WT-V5 (Figure 2a,b; Figure S1). These results suggest that both variants impede NatA complex formation, which ultimately is likely to cause reduced NatAmediated Nt-acetylation in vivo.
The catalytic activity of NAA10-D10G-V5 and NAA10-L11R-V5 was investigated through in vitro Nt-acetylation assays (Figure 2c,d). In these assays, SESS24 represented a canonical NatA substrate, while EEEI24 was used as a substrate for monomeric NAA10, which preferentially Ntacetylates acidic N-termini in vitro. Therefore, the measured catalytic activity against SESS24 and EEEI24 was normalised to the amount of NAA15 and NAA10-V5, respectively. The Nt-acetylation assays revealed that NAA10-D10G-V5 displayed normal NatA catalytic activity, whereas the monomeric NAA10-D10G-V5 NAT activity was reduced by approximately two-fold as compared to NAA10-WT-V5 ( Figure 2c). In contrast, NAA10-L11R-V5 demonstrated the opposite effect, the NatA catalytic activity being reduced compared to NAA10-WT-V5, while the monomeric NAT activity was not affected (Figure 2d).
In summary, the experimental evidence suggests that both variants hamper formation of the NatA complex but have different effects on the enzymatic activity of NAA10. NAA10-D10G-V5 showed normal activity in the formed NatA complex but impaired NAA10 monomeric activity, while NAA10-L11R-V5 showed impaired NatA activity but normal monomeric NAA10 activity. The amino acids D10 and L11 are highly conserved in the NAA10 sequences (conservation is coloured red). (b) NAA10 (green) and NAA15 (blue) in the human NatA structure (PDB ID: 6C9M) [21] bound with ligand IP 6 (magenta). The human NatA structure was superimposed with the substrate (SASE, yellow sticks; CoA, pink sticks) from the S. pombe NatA structure (PDB ID: 4KVM) [15]. A zoomed-in view (right) shows D10 (purple) and L11 (orange) and the interactions they form with surrounding amino acids. , was compared to NAA10-WT-V5 through in vitro Nt-acetylation assays. Negative controls included reactions without peptide and with β-gal-V5 IP. The measured activities towards the canonical NatA complex substrate SESS24 and the in vitro monomeric NAA10 substrate EEEI24 were normalised to the amount of NAA15 and NAA10-V5 used in the reaction, respectively. The Western blot analyses and Nt-acetylation assays shown are representative of three independent experiments. The Nt-acetylation assays were performed with three technical replicates.

Discussion
Functional characterisations of different NAA10 variants identified in individuals have revealed that distinct variants can affect the biological roles of NAA10 differently. We report here the functional characterisation of two de novo NAA10 variants, p.(D10G) and p.(L11R). NAA10 p.(D10G) and NAA10 p.(L11R) were previously identified in a boy and a girl, respectively, with phenotypes The catalytic activity of NAA10-D10G-V5 and NAA10-L11R-V5 was investigated through in vitro Nt-acetylation assays (Figure 2c,d). In these assays, SESS 24 represented a canonical NatA substrate, while EEEI 24 was used as a substrate for monomeric NAA10, which preferentially Nt-acetylates acidic N-termini in vitro. Therefore, the measured catalytic activity against SESS 24 and EEEI 24 was normalised to the amount of NAA15 and NAA10-V5, respectively. The Nt-acetylation assays revealed that NAA10-D10G-V5 displayed normal NatA catalytic activity, whereas the monomeric NAA10-D10G-V5 NAT activity was reduced by approximately two-fold as compared to NAA10-WT-V5 (Figure 2c). In contrast, NAA10-L11R-V5 demonstrated the opposite effect, the NatA catalytic activity being reduced compared to NAA10-WT-V5, while the monomeric NAT activity was not affected (Figure 2d).
In summary, the experimental evidence suggests that both variants hamper formation of the NatA complex but have different effects on the enzymatic activity of NAA10. NAA10-D10G-V5 showed normal activity in the formed NatA complex but impaired NAA10 monomeric activity, while NAA10-L11R-V5 showed impaired NatA activity but normal monomeric NAA10 activity.

Discussion
Functional characterisations of different NAA10 variants identified in individuals have revealed that distinct variants can affect the biological roles of NAA10 differently. We report here the functional characterisation of two de novo NAA10 variants, p.(D10G) and p.(L11R). NAA10 p.(D10G) and NAA10 p.(L11R) were previously identified in a boy and a girl, respectively, with phenotypes including developmental delay and brain malformations [40].
Bioinformatic analyses of the NAA10 sequence and structure revealed that D10 and L11, as well as the amino acids they interact with, are highly conserved in NAA10 orthologues (Figure 1a,b), indicating that D10 and L11 are important for NAA10 stability or function. The missense variants NAA10 p.(D10G) and NAA10 p.(L11R) are likely to disrupt the electrostatic interaction between D10 and R4 and the hydrophobic pocket that L11 occupies, respectively. Thus, we hypothesised that these missense variants may cause structural changes that impede the NAA10 active site as well as the interactions between the NAA10 α1 helix and NAA15.
Biochemical experiments were performed to investigate how the variants impacted NatA complex formation and enzymatic activity. Immunoprecipitation of both variants revealed that they had reduced capacity to bind NAA15 compared to NAA10-WT-V5 (Figure 2a,b). This was probably caused by disruption of important interactions between the NAA10 α1 helix and NAA15 [15]. Furthermore, in vitro Nt-acetylation assays showed that NAA10-L11R-V5 had reduced NatA activity and NAA10-D10G-V5 had equal NatA activity compared to NAA10-WT-V5 (Figure 2c,d). Since both variants impair NatA complex formation, it is likely that they both result in less Nt-acetylation of NatA substrates at the ribosome in patient cells. Additionally, NAA10 p.(L11R) probably showed a stronger decrease in Nt-acetylation of NatA substrates than NAA10 p.(D10G) because of its debilitated NatA function, even in the formed NatA complex (Figure 2d).
The monomeric NAA10 NAT activity was only affected by the NAA10 p.(D10G) variant and not NAA10 p.(L11R). This suggests that NAA10 p.(D10G) may affect the NatA-independent roles of NAA10, resulting in diverse downstream effects. To date, no in vivo NAT substrates for monomeric NAA10 have been identified. However, NAA10 has been reported to catalyse the lysine acetylation of several proteins as well as regulating proteins in a non-catalytic manner [3]. Whether these NAT-independent roles of NAA10 are affected by the NAA10 p.(D10G) or NAA10 p.(L11R) variants remains unknown.
The NAA10 p.(D10G) and NAA10 p.(L11R) variants were previously reported by Cheng et al. in an international cohort of individuals with NAA10 variants [40]. In this cohort study, both variants were found to cause a decreased in vitro thermostability of NatA compared to NatA WT, which is in accordance with our IP evidence showing that both variants have a decreased binding affinity for NAA15 in vivo. Furthermore, Cheng et al. tested the in vitro hNatA catalytic activity of NAA10 variants purified from insect Sf9 cells. While both NAA10 p.(D10G) and NAA10 p.(L11R) were found to have reduced NatA activity, NAA10 p.(D10G) was more strongly affected than NAA10 p.(L11R). This contradicts our findings in which NAA10 p.(D10G) NatA activity was unaffected in the formed NatA complex. One explanation for this discrepancy could lie in the nature of the experiment; we tested NatA immunoprecipitated from human cells in which potential binding partners and/or post-translational modifications not present in insect cells could influence enzymatic activity. HYPK is a known binding partner of NatA. In fact, Cheng et al. found that the NatA activity of NAA10 p.(L11R) was not reduced when bound to HYPK, which highlights the intricacies of determining the impact of NAA10 variants on the NatA machinery.
Since males only have one X-chromosome, they are generally more severely affected by NAA10 variants than heterozygous females. The boy harbouring NAA10 p.(D10G) died during infancy and had similar clinical features to males harbouring NAA10 p.(S37P)-a lethal condition originally called Ogden syndrome [32]. Similar to NAA10 p.(D10G) and p.(L11R), NAA10 p.(S37P) has been found to impair NatA complex formation and cause reduced Nt-acetylation of some NatA substrates in Ogden-syndrome patient fibroblasts and B-cells. Female carriers of NAA10 p.(S37P) had no severe symptoms, which is possibly explained by their X-chromosome inactivation pattern favouring the WT allele [33]. The impact of X-chromosome inactivation skewing on the clinical manifestations in the female carrying NAA10 p.(L11R) is unclear, as it has not been experimentally tested. Recently, a de novo NAA10 p.(H16P) variant was identified in a young female with severe ID and DD [39]. The NAA10 p.(H16P) variant was located in the NAA10 α1 helix and functional studies revealed that this variant also hampered NatA complex formation. Furthermore, the NAA10 p.(H16P) enzymatic activity resembled that of NAA10 p.(L11R), in that the NatA activity was reduced but the monomeric NAA10 NAT activity was intact. Another NAA10 variant, p.(N101K), was recently described in two unrelated girls with DD and hemihypertrophy [43]. The NAA10 p.(N101K) variant was shown to completely abolish NatA complex formation and NatA enzymatic activity, while its monomeric NAA10 NAT activity was unaffected. NAA10 p.(L11R) may thus have a similar disease mechanism as NAA10 p.(H16P) and NAA10 p.(N101K), in which the phenotypes are caused by impaired NatA Nt-acetylation at the ribosomes.
In conclusion, NAA10 p.(D10G) and NAA10 p.(L11R) appear to be causative of disease through decreased NatA complex formation and consequently reduced NatA-mediated Nt-acetylation of NatA substrates. Since NatA has a large substrate pool, approximately 40% of the human proteome [3], it is challenging to define all the downstream effects this can cause. NAA10 p.(D10G) may additionally impede the individual roles of NAA10 in the cell. More research is needed in order to unravel the exact disease mechanisms involved with pathogenic NAA10 variants and understand their effect on NAT, KAT and the non-catalytic functions of NAA10.

In Vitro N-terminal (Nt)-Acetylation Assays
In vitro Nt-acetylation assays were performed as described [47]. Triplicate reactions were run in each assay, and contained the following: 10 µL immunoprecipitated enzyme, 200 µM synthetic oligopeptide SESS 24 (SESSSKSRWGRPVGRRRRPVRVYP) or EEEI 24 (EEEIAALRWGRPVGRRRRPVRVYP) (>95% purity, BioGenes), 50 µM [14C]-Ac-CoA (PerkinElmer, Waltham, MA, US) and acetylation buffer to a final volume of 25 µL. Negative control reactions contained either immunoprecipitated β-gal-V5 or no substrate peptide. Reactions were incubated at 37 • C for 30 min on a thermoshaker (Eppendorf, Hamburg, Germany). To measure product formation, the reaction was quenched by transferring 23 µL of the reaction mixture onto P81 phosphocellulose filter discs (Millipore, Burlington, MA, US). The filter discs were washed three times in 10 mM HEPES buffer (pH 7.4) and dried before adding them to 5 mL Ultima Gold F scintillation mixture (PerkinElmer). Product formation was measured by a TriCarb 2900TR Liquid Scintillation Analyzer (PerkinElmer). The two variants were tested in independent setups compared to the WT enzyme, thus explaining some variability in the relative activity of WT NAA10 towards different substrates.

Conflicts of Interest:
The funders had no role in the design of the study; in the collection, analysis or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Abbreviations
Nt N-terminal NAT N-terminal acetyltransferase KAT lysine acetyltransferase DD developmental delay ID intellectual disability IP immunoprecipitation