BRIEF RESEARCH REPORT published: 26 May 2021 doi: 10.3389/fnins.2021.618098 The Novel KIF1A Missense Variant (R169T) Strongly Reduces Microtubule Stimulated ATPase Activity and Is Associated With NESCAV Syndrome Cinthia Aguilera 1† , Stefan Hümmer 2,3† , Marc Masanas 4 , Elisabeth Gabau 5 , Miriam Guitart 1 , A. Arockia Jeyaprakash 6 , Miguel F. Segura 4 , Anna Santamaria 7* and Anna Ruiz 1* 1 Genetics Laboratory, UDIAT-Centre Diagnòstic, Parc Taulí Hospital Universitari, Institut d’Investigació i Innovació Parc Taulí I3PT, Universitat Autònoma de Barcelona, Sabadell, Spain, 2 Translational Molecular Pathology, Vall d’Hebron Research Edited by: Institute (VHIR), Universitat Autònoma de Barcelona (UAB), Barcelona, Spain, 3 Spanish Biomedical Research Network Filomena Pirozzi, Centre in Oncology (CIBERONC), Madrid, Spain, 4 Group of Translational Research in Child and Adolescent Cancer, Vall Seattle Children’s Research Institute, Hebron Research Institute (VHIR), Universitat Autònoma de Barcelona (UAB), Barcelona, Spain, 5 Paediatric Unit, Parc Taulí United States Hospital Universitari, Institut d’Investigació i Innovació Parc Taulí I3PT, Universitat Autònoma de Barcelona, Sabadell, Spain, Reviewed by: 6 Wellcome Centre for Cell Biology, University of Edinburgh, Edinburgh, United Kingdom, 7 Cell Cycle and Cancer Laboratory, Shinsuke Niwa, Group of Biomedical Research in Urology, Vall Hebron Research Institute (VHIR), Universitat Autònoma de Barcelona (UAB), Tohoku University, Japan Barcelona, Spain Henna Tyynismaa, University of Helsinki, Finland KIF1A is a microtubule-dependent motor protein responsible for fast anterograde *Correspondence: Anna Santamaria transport of synaptic vesicle precursors in neurons. Pathogenic variants in KIF1A have anna.santamaria@vhir.org been associated with a wide spectrum of neurological disorders. Here, we report a Anna Ruiz aruizn@tauli.cat patient presenting a severe neurodevelopmental disorder carrying a novel de novo † These authors have contributed missense variant p.Arg169Thr (R169T) in the KIF1A motor domain. The clinical features equally to this work present in our patient match with those reported for NESCAV syndrome including severe developmental delay, spastic paraparesis, motor sensory neuropathy, bilateral optic Specialty section: This article was submitted to nerve atrophy, progressive cerebellar atrophy, epilepsy, ataxia, and hypotonia. Here, we Neurogenomics, demonstrate that the microtubule-stimulated ATPase activity of the KIF1A is strongly a section of the journal Frontiers in Neuroscience reduced in the motor domain of the R169T variant. Supporting this, in silico structural Received: 16 October 2020 modeling suggests that this variant impairs the interaction of the KIF1A motor domain Accepted: 23 March 2021 with microtubules. The characterization of the molecular effect of the R169T variant on Published: 26 May 2021 the KIF1A protein together with the presence of the typical clinical features indicates its Citation: causal pathogenic effect. Aguilera C, Hümmer S, Masanas M, Gabau E, Guitart M, Keywords: NESCAV syndrome, KIF1A, kinesin, microtubule, motility, ATPase Jeyaprakash AA, Segura MF, Santamaria A and Ruiz A (2021) The Novel KIF1A Missense Variant (R169T) Strongly Reduces INTRODUCTION Microtubule Stimulated ATPase Activity and Is Associated With The kinesin superfamily of proteins (KIFs) are microtubule-dependent molecular motors that NESCAV Syndrome. participate in the transport of membrane vesicles, organelles, protein complexes, and mRNAs along Front. Neurosci. 15:618098. microtubules, thus playing important roles including mitosis, meiosis, and transport of cellular doi: 10.3389/fnins.2021.618098 cargo such as axonal transport (Hirokawa et al., 1989). KIF1A belongs to the kinesin-3 family of Frontiers in Neuroscience | www.frontiersin.org 1 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T kinesins and is a neuron-specific protein composed of an along microtubules, resulting in a lower impairment of synaptic N-terminal motor domain (MD) (1–361 aa), which contains the vesicle anterograde transport and possibly only affecting long ATPase activity and the microtubule-binding domains. These axons. Alternatively, it has been suggested that a redundancy in domains are followed by a short-strand “neck linker” (353– the expression of kinesins in affected neuronal populations can 361 aa) and a neck coil region (NC) (365–397 aa) which are compensate KIF1A variants with milder effect (Hirokawa and key regulatory elements for kinesin-3 motor dimerization and Tanaka, 2015; Pennings et al., 2020). However, a recent study has processive motility. The coiled-coil domains [CC1 (429–462 aa), demonstrated that some spastic paraplegia pathogenic variants CC2 (622–681 aa), and CC3 (801–822 aa)] together with the (V8M, A255V, and R350G) confer a gain-of-function activity to forkhead-associated domain (516–572 aa) are involved in the the KIFA protein resulting in a hyperactivation of the kinesin autoregulation of the motor activity and in the dimerization of motor domain (Chiba et al., 2019). Thus, the reason why different the protein. The pleckstrin homology domain (1676–1774 aa) is variants cause different types of diseases still remains unclear. implicated in cargo binding (Klopfenstein et al., 2002; Hammond Here, we present the identification of a patient affected by a et al., 2009; Huo et al., 2012). KIF1A exists in a dimeric inactive severe neurodevelopment disorder who is a carrier of a novel state that is maintained by autoinhibitory mechanisms and are de novo missense variant in the motor domain of KIF1A. The activated upon cargo binding, enabling KIF1A to transport variant p.Arg169Thr (R169T) has not been described before in synaptic vesicle precursors along microtubules (Okada et al., other patients with ID or spastic paraplegia and is not present in 1995; Lee et al., 2003; Hammond et al., 2009). the gnomAD database of control individuals. In silico prediction Pathogenic variants in KIF1A have been linked to four tools suggest a deleterious effect of the variant on the protein. clinical different disorders: (i) neuropathy, hereditary sensory, In vitro analysis of the R169T variant revealed a strong inhibition type IIC (HSNIIC, MIM # 614213); (ii) Spastic Paraplegia of microtubule-stimulated ATPase activity. Moreover, structure- 30, autosomal recessive (SPG30, MIM # 610357); (iii) Spastic based modeling predicts that R169T variant may directly interfere Paraplegia 30, autosomal dominant (SPG30, MIM # 610357); and with the interaction of KIF1A motor domain with microtubules. iv) neurodegeneration and spasticity with or without cerebellar atrophy or cortical visual impairment (NESCAV syndrome, NESCAVS, MIM # 614255, formerly known as mental retardation MATERIALS AND METHODS autosomal dominant 9). The NESCAV syndrome always occurs as a consequence The study has been approved by the institutional Ethics of de novo missense variants located in the motor domain. Committee of Institut d’Investigació i Innovació Parc Taulí I3PT The clinical features include moderate to severe intellectual (CEIC 2019/514), and the corresponding informed consent has disability (ID), language and motor delay, hypotonia, spastic been obtained from the parents. paraparesis, hyperreflexia, postnatal microcephaly, and peripheral neuropathy. MRI shows optic nerve atrophy and Whole-Exome Sequencing and Variant varying degrees of brain atrophy, the cerebellum being the most Detection severely affected. It is thought that the variants leading to this Trio WES of the patient and his parents was performed using the clinical presentation disrupt the motility of the KIF1A protein SureSelect Human All Exon V5+UTR kit (Agilent Technologies, along the microtubules (Esmaeeli Nieh et al., 2015; Lee et al., Santa Clara, United States). Sequencing was performed at the 2015). National Center of Genomic Analysis (CNAG-CRG, Barcelona, Functional studies performed on hippocampal neurons show Spain) on an Illumina HiSeq 2000 platform (Illumina, San a significant reduced distal localization of mutant motor domain Diego, CA, United States), producing 2 × 100 nucleotides constructs harboring the de novo missense variants T99M, A202P, paired end reads. Raw reads were mapped to the human S215, R216P, and E253K compared with WT motor domain in reference genome (hg19) using the BWA aligner (Li and Durbin, neurites (Lee et al., 2015). In addition, experiments performed in 2010) and subsequently processed using the GATK pipeline SH-SY5Y cells demonstrated that the de novo missense variants in order to remove PCR duplicates and perform base quality C92R, D248E, and P305L caused an accumulation of mutated score recalibration. Variant discovery was performed using the protein preferentially in the cell body instead to the tips of Haplotype Caller tool and following the best practices for exome neurites, where wild-type KIF1A is located (Kaur et al., 2020). sequencing variant discovery of GATK (DePristo et al., 2011). Moreover, in vitro gliding assays showed loss of motility for de Only variants with a called genotype and genotype quality ≥ 20 novo missense C92R, T99M, R216C, D248E E253K, and P305L in all the trio members were considered. mutant motor domains (Esmaeeli Nieh et al., 2015; Kaur et al., Finally, all variants were annotated using ANNOVAR (Wang 2020). The equimolar mix of WT and E253K showed an 8- et al., 2010). All exome variants were checked against a de novo fold reduction in microtubule gliding velocities, suggesting a model of inheritance, and variants were filtered for a population dominant negative effect (Esmaeeli Nieh et al., 2015). allele frequency of less than 1/10,000 in the gnomAD database Interestingly, alternative missense variants in the motor (Karczewski et al., 2020) and a predicted deleterious effect on domain have been demonstrated to be responsible for the the protein (nonsense, frameshift, splicing, and missense variants dominant and recessive spastic paraplegia phenotype not were prioritized). The impact of missense variants was assessed associated with ID. It has been proposed that variants causing using several in silico tools including PolyPhen-2 (Adzhubei this phenotype may have a milder effect on the motility of KIF1A et al., 2010), Mutation assessor (Reva et al., 2007), PROVEAN Frontiers in Neuroscience | www.frontiersin.org 2 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T (Choi and Chan, 2015), MutPred (Pejaver et al., 2017), CADD SDS-PAGE gel and transferred onto a PVDF membrane (Sigma- (Rentzsch et al., 2019), Primate AI (Sundaram et al., 2018) and Aldrich). The membrane was blocked in 5% BSA (Sigma-Aldrich) REVEL (Ioannidis et al., 2016). and incubated with the Anti-KIF1A antibody (diluted 1:1,000, Abcam ab180153). Plasmids pMA-RQ KIF1A motor domain WT (3–362 aa) and R169T In vitro ATPase Assay were obtained from GeneArt Gene synthesis service (Thermo The in vitro ATPase assay was performed with the Kinesin Fisher Scientific, Massachusetts, United States). pcDNA3.1(+) ATPase Endpoint Biochem kit (Cytoskeleton Inc., Denver, KIF1A-C-Myc was obtained from GenScript Biotech (Clone United States). The concentration of KIF1A WT and R169T ID OHu22637). To obtain the R169T variant, site-directed motor domains was adjusted to 0.5 µg/µl, then a dilution mutagenesis was performed using the QuikChange II Site- to a final concentration of 0.03 µg/µl was performed using directed mutagenesis kit (Agilent Technologies, Santa Clara, Kinesin Reaction Buffer. To test the ATPase activity, the indicated United States) following the manufacturer’s instructions. amounts of the diluted WT and R169T motor of domains, Mutagenesis primers are listed in Supplementary Table 1. ranging from 0.0078 to 0.24 µg, were added to each reaction. For the basal ATPase activity, Kinesin Reaction Buffer without Cloning and Production of Recombinant microtubules was added to a final volume of 30 µl. For the microtubule-stimulated ATPase activity, 10 µl of taxol-stabilized KIF1A R169T and WT Motor Domains microtubules (0.2 mg/ml) and Kinesin Reaction Buffer to a The motor domains of KIF1A WT and R169T (3–362 aa, final volume of 30 µl were added. The ATPase reaction was NP_001230937.1) were obtained from the GeneArt Gene started by adding 10 µl of 100 mM ATP PIPES pH 7.0 synthesis service cloned in the pMA-RQ plasmid. The WT (Sigma Aldrich, Missouri, United States) and incubating 5 min and R169T motor domains were subcloned into a pQE-80L at room temperature. To terminate the reaction, 70 µl of expression vector with a hexa-histidine tag, kindly provided CytoPhos reagent was added and incubated for 10 min at room by Dr. Thomas U. Mayer. The constructs were transformed temperature. Absorbance at 650 nm was measured by Epoch in the BL21-Codon Plus RIL cells (Agilent technologies), and Microplate Spectrophotometer (BioTeK, Agilent Technologies). protein expression was induced with 1 mM IPTG (isopropyl In order to calculate the activity, the absorbance of the reaction β-d-1-thiogalactopyranoside), at 18◦ C overnight. The expression with microtubules was subtracted from the absorbance of and solubility of WT and R169T motor domains were tested the reaction without microtubules. ATPase activity was tested by western blot, staining the PVDF (polyvinylidene difluoride) in triplicate, and the estimation of the amount of inorganic membrane with Naphthol Blue and using a penta-his antibody phosphate (Pi) released through the ATPase activity was (included in QIAexpress Ni-NTA Fast Start kit). Protein determined by using a Pi standard curve. For experiments with purification was performed with the QIAexpress Ni-NTA Fast increasing concentrations of microtubule form (0–4.375 µM), Start kit (Qiagen, Hilden, Germany) following the manufacturer’s 0.015 µg/µl of WT motor domain and 0.06 µg/µl of R169T were instructions. After purification, a western blot and Naphthol used. Titration curves were performed in triplicate, and activity Blue staining of the PVDF membrane was carried out in (nmol min−1 ) was blotted against microtubule concentration order to confirm the presence of the purified WT and R169T (µM). Double reciprocal analysis of the ATPase activity vs. motor domains. Finally, the recombinant proteins were dialyzed microtubule concentration was performed and data were fitted by overnight at 4◦ C against a buffer containing 20 mM Tris–HCl pH linear regression. Km and Vmax were determined as intersections 7.4, 300 mM NaCl, 10 mM β-mercaptoethanol, 10% de glycerol, 1 with the X = K−1m and Y = Vmax axis, respectively, according and 150 µM ATP. to the Lineweaver Burk equation ( V = VKmax 1 m 1 · [S] 1 + Vmax ). Kcat Vmax Overexpression of KIF1A in the SH-SY5Y was determined by Kcat = [Total enzyme concentration] . Cell Line The human neuroblastoma cancer cell line SH-SY5Y (bone Structural Modeling of the KIF1A R169T marrow) was purchased from the American Type Culture Missense Variant Collection (ATCC, CRL-2266). Cells were grown at 37◦ C The R169T mutation was modeled in the published KIF1A motor under a humidified atmosphere (5% CO2 ) conditions in domain crystal structure (2HFX) using PyMOL, a molecular Iscove’s modified Dulbecco’s Medium (Thermo Fisher Scientific) graphics program (The PyMOL Molecular Graphics System, supplemented with 10% fetal bovine serum (Sigma-Aldrich) Version 1.2r3pre, Schrödinger, LLC). and 1% of Insulin-Transferrin-Selenium Supplement. A total of 520,000 cells/well were transfected with pcDNA3.1 (+) empty, pcDNA3.1 (+) KIF1A-C-myc WT, and KIF1A-C-myc R169T RESULTS using Lipofectamine 2000 (Thermo Fisher Scientific), following the manufacturer’s instructions. Cells were homogenized in Clinical Description of the Patient RIPA lysis buffer (Thermo Fisher Scientific), and total protein The proband is a 10-year-old boy born at term from non- content was quantified using the DC Protein Assay Kit (Bio- consanguineous parents. Birth weight was 2,750 g. No perinatal Rad). Thirty micrograms of protein was separated on 4–12% problems were reported, except for gastroesophageal reflux. At Frontiers in Neuroscience | www.frontiersin.org 3 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T FIGURE 1 | MRI findings and pictures of the patient. (A) MRI showing progressive cerebellar atrophy. (B) Patient at 1 year of age (top left), at 1 year and 9 months of age (top right) and at 8 years of age (in the bottom). the age of 10 months, he presented global development delay, ACMG guidelines (Richards et al., 2015). The variant has not microcephaly, hypotonia, feeding difficulties, stereotypies, and an been described before and was submitted to the ClinVar database abnormal electroencephalogram (EEG). MRI showed a delayed (ref. VCV000666256). myelination. Genetic analysis included Angelman syndrome testing (15q11-q13 methylation pattern and UBE3A sequencing), Functional Analysis of the R169T Variant karyotype, fragile X syndrome testing, subtelomeric MLPA The KIF1A R169T Variant Does Not Alter the Protein (SALSA MLPA P070, MRC Holland), and 60K array Comparative Stability Genomic Hybridization (aCGH) (Agilent Technologies) with Missense variants can affect the protein folding as well as protein normal results. At the age of 5 years, the patient developed stability as a consequence of changes in the physiochemical epilepsy and MRI scans showed partial bilateral optic atrophy and characteristics or conformational constraints. It has been progressive cerebellar atrophy (Figure 1A). Currently, the patient presents severe ID, speech impairment, postnatal progressive microcephaly (-2SD), hypotonia, partial occipital epilepsy, TABLE 1 | Clinical features of the patient according to what is described for unsteady gait, feeding problems (dysphagia), hyperreflexia, NESCAV syndrome. ataxia, peripheral motor sensory neuropathy, and short stature. Clinical features associated Patient with c.506G > C, p.Arg169Thr de The behavioral phenotype includes apparent happy demeanor, with NESCAV syndrome novo missense variant easily excitable personality, and stereotypies (Figure 1B and Moderate to severe + Table 1). developmental delay/ID Severe motor delay + Ataxia + Identification of a de novo R169T Variant Peripheral neuropathy + in the KIF1A Gene Axial hypotonia + Trio-WES analysis identified a total of three variants after Spastic paraparesis + filtering the exome data according to a de novo model of Hyperreflexia + inheritance, a population allele frequency of less than 1/10,000, Microcephaly + and a predicted deleterious effect on the protein (Supplementary Optic nerve atrophy + Table 2). A novel de novo missense variant, c.506G > C, Abnormal eye movements + Mild to moderate language delay + p.Arg169Thr (R169T), in the KIF1A gene (NM_001244008.2) Epilepsy + was selected as the best candidate. De novo variants in the KIF1A Epileptic abnormalities on EEG + gene are associated with NESCAV syndrome, whose clinical Metatarsus adductus + features match with those described in our patient (Table 1; MRI Esmaeeli Nieh et al., 2015; Lee et al., 2015). The variant is not Brain atrophy – present in the gnomAD database. Several bioinformatic tools Diminished cerebral white matter – were used to predict the functional effect of the variant on the Severe and progressive + protein, all of them suggesting a strong damaging effect (Table 2). cerebellar atrophy The variant was classified as likely pathogenic following the +, present; –, not present. Frontiers in Neuroscience | www.frontiersin.org 4 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T TABLE 2 | In silico analysis of the missense variant p.Arg169Thr. Structural Modeling of the KIF1A R169T In silico tool Prediction Score Missense Variant The KIF1A kinesin motor domain is composed of a single PROVEAN Deleterious –5.36 globular catalytic core (1–361 aa) that contains domains for both PolyPhen-2 Probably damaging 0.991 ATPase activity and microtubule binding. The amino acid R169 Mutation assessor Medium 3.29 is highly conserved throughout evolution, and it is located in MutPred Deleterious 0.833 the β5-loop 8 corresponding to one of the three microtubule- CADD Deleterious 26.2 binding regions (together with loop 11 and loop 12) (Figure 3A; PrimateAI Damaging 0.8787 Kikkawa et al., 2000; Scarabelli et al., 2015). The variant R169T REVEL Pathogenic 0.8569 eliminates an arginine that has a positively charged side chain and introduces a threonine that has a polar uncharged side chain. As shown in Figure 3B, R169 is involved in electrostatic reported that over 80% of all disease-associated missense variants interactions with E417 and E420 of β-tubulin. Mutating R169T are amino acid substitutions that affect protein stability (Wang likely weakens the electrostatic interaction between the KIF1A and Moult, 2001; Zhang et al., 2012). In the absence of a cell protein and β-tubulin, suggesting a reduced KIF1A interaction line model carrying the variant R169T, we ectopically expressed with microtubules, which might ultimately cause the observed the WT and mutant KIF1A forms in SH-SY5Y neuroblastoma reduction in the microtubule-stimulated ATPase activity. cells and determined protein stability by western blot. The To deepen into the contribution of the β5-loop 8 for results show that the R169T and WT constructs are expressed at microtubule binding in KIF1A, we superposed the structure of similar levels and had the same electrophoretic mobility, thereby the microtubule-bound KIF1A (PDB: 2HXF) onto microtubule- indicating that the ectopically expressed variant R169T does not bound Kinesin-1 (PDB: 3J8Y). The substitution of H156A mildly affect the protein stability (Figure 2A). affected Km and Kcat (1.62 ± 0.43 and 1.12 ± 0.02, respectively) (Woehlke et al., 1997). On the contrary, the substitution R169T on KIF1A strongly affects Km and Kcat (Figure 2F). R169 and R169T Missense Variant Impairs KIF1A ATPase R167 into the β5-loop 8 of KIF1A make more interactions with Activity microtubules as compared to the equivalent residues H156 and In order to assess the effect of the variant on the functionality S154 in Kinesin-1 (Figure 3C). In KIF1A, both R169 and R167 of the KIF1A motor domain, we purified the wild-type and make hydrogen bonding interactions with tubulin residues, E420 mutant recombinant KIF1A motor-domain proteins (3–362 aa) and E417. However, equivalent residues in Kinesin-1, H156, and (Supplementary Figure 1) and performed an ATPase assay. The S154 are not within hydrogen bonding distance to E420 and inorganic phosphate (Pi) generated at different concentrations E417 tubulin amino acids (Figure 3C). In addition, a 67.4-Å2 of purified WT and R169T motor domains was measured surface area of R169 is contributed to the total surface area buried following induction of microtubule-stimulated ATPase activity. upon KIFA–tubulin complex formation, which is 5% of the total We found that while the release of Pi increased in function area buried (1,324 Å2 ) at the binding interface of β-tubulin and of the concentration of the WT motor domain, there was KIF1A. Nevertheless, only 23.3 Å2 is buried in Kinesin-1, which a strong reduction in the amount of Pi generated by the is 1.8% of the total area buried (1,271 Å2 ) at the interface between R169T motor domain. The microtubule-stimulated ATPase β-tubulin and kinesin-1 (not shown). Together, these results rate of the R169T motor domain was about 180-fold lower show that the contribution of H156 for microtubule binding than the motor domain of WT KIF1A (8.3 nmol of ATP in Kinesin-1 is relatively minor as compared to that of R169 hydrolyzed/min/mg vs. 1466.6 nmol of ATP hydrolyzed/min/mg; in KIF1A, likely explaining why mutating R169 in KIF1A has Student’s t-test, p-value 0.0095). Thus, the ATPase activity of the a stronger effect as compared to that of H156 on Kinesin-1 KIF1A motor domain is severely affected by the variant R169T (Woehlke et al., 1997). (Figures 2B–D). To test if this defect is caused by impaired stimulation of the ATPase activity by microtubules, we determined the DISCUSSION ATPase activity in KIF1A WT and R169T with different amounts of microtubules. In the case of the R169T mutant form, KIF1A encodes a neuron-specific motor protein that plays an ATPase activity can only be achieved in the presence of high important role in anterograde axonal transport of synaptic concentrations of microtubules (i.e., 2.18 µM) (Figure 2E). vesicle precursors. Pathogenic variants in the KIF1A gene Determination of the Km and Kcat values revealed an 11- have been associated with a wide spectrum of neurological fold increase in Km (0.28 ± 0.14 µM vs. 3.16 ± 1.41 phenotypes (Nemani et al., 2020) ranging from recessive µM) and 8-fold decrease in Kcat (11.66 ± 2.97 s−1 vs. hereditary sensory neuropathy-type IIC (HSNIIC) to autosomal 1.48 ± 0.27 s−1 ) of the R169T mutant compared with the recessive or dominant hereditary spastic paraplegia 30 (SPG30) WT motor domain (Figures 2E,F). Therefore, these results and to the most severe NESCAV syndrome (NESCAVS). There indicate that the variant R169T strongly affects the microtubule- are approximately more than 80 variants located in the motor stimulated ATPase activity, which results in a reduction in the domain and associated with KIF1A-related disorders (Kaur et al., maximal ATPase rate. 2020). The differences in disease severity between NESCAVS, Frontiers in Neuroscience | www.frontiersin.org 5 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T FIGURE 2 | Functional assays carried out with the R169T variant. (A) Overexpression of full-length KIF1A WT and R169T in SH-SY5Y neuroblastoma cells. Cell lysates were harvested 96 h post-transfection. Thirty micrograms of protein corresponding to SH-SY5Y cells transfected with pcDNA3.1empty , pcDNA3.1 KIF1AWT , and pcDNA3.1 KIF1AR169T vectors loaded. Actin was used as a loading control. (B) Naphthol Blue staining of the PDVF membrane showing that the same amount of protein corresponding to WT and R169T has been used for the ATPase assay. (C) ATPase activity assay diagram. (D) Microtubule-stimulated ATPase activity of the WT and R169T motor domains. A range of 0.0078–0.24 µg of recombinant protein was used in the assay. To start the ATPase reaction, 10 µl of taxol-stabilized microtubules (0.2 mg/ml) and 10 µl of ATP 100 mM ATP PIPES pH 7.0 were added to each well and incubated 5 min at room temperature. The readout at 650 nm was performed after a 10-min incubation of the reactions with 70 µl of CytoPhos termination reagent. (E) ATPase activity in the presence of increasing concentrations of microtubules. A range of 0–4.375 µM of microtubules was used. Data were fitted by linear regression. Km and Vmax were determined according to the Lineweaver Burk equation. Kcat was determined by Kcat = Vmax /[total enzyme concentration]. Pi, inorganic phosphate; MTs, microtubules. (F) Km and Kcat values of KIF1A motor domain (MD) WT and R169T. Frontiers in Neuroscience | www.frontiersin.org 6 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T FIGURE 3 | Conservation analysis and structural modeling of the R169T variant. (A) Multiple-sequence alignment showing the conservation of R169 residue in KIF1A across evolution. The alignment has been performed using HomoloGene (https://www.ncbi.nlm.nih.gov/homologene). L7, loop 7; L8, loop 8; MB1, microtubule-binding region 1. (B) Schematic representation of KIF1A bound to the alpha/beta tubular heterodimer. Insets show close-up views of (left) the electrostatic interaction mediated by KIF1A R169 with β-tubulin residues E417 and E420 and (right) the loss of electrostatic interaction caused by KIF1A R169T mutation. α-Tubulin is shown in green, β-tubulin in cyan, and KIF1A in pink. (C) Superposed structure of microtubule-bound Kinesin-1 (PDB: 3J8Y) onto microtubule-bound KIF1A (PDB: 2HXF). with a severe neurological involvement, and the relative mild R350W are found in dominant SPG while variants G102D and HSNIIC and both recessive and dominant SPG30 forms is R167C cause NESCAVS and R350G recessive SPG. thought to be due to the effect of the variant on protein transport Here, we report a novel de novo missense variant (R169T) by the KIF1A protein. located in the motor domain of the KIF1A protein. The patient It is hypothesized that mild transport defects would be presents the typical features of NESCAV syndrome, severe responsible for the hereditary spastic paraparesis phenotype, developmental delay, ataxia, spasticity, progressive cerebellar whereas severe defects in transport would be responsible for the atrophy, peripheral neuropathy, and optic bilateral atrophy NESCAVS phenotype (Pennings et al., 2020). However, a recent (Table 1 and Figure 1). First, we showed that the full-length study has shown that some missense variants associated with KIF1A WT and R169T mutants were expressed in similar levels in SPG30 (V8M, A225V, and R350G) located in the motor domain human neuronal cells and with identical electrophoretic mobility, produce a hyperactivation of KIF1A motor activity in vitro thereby suggesting no major alterations in protein stability. (10–20-fold higher than the WT), resulting in an increased However, we cannot exclude the possibility that the endogenous anterograde transport of synaptic vesicles in Caenorhabditis protein in the patient might be unstable. In fact, a computational elegans (Chiba et al., 2019). analysis of 22 KIF1A and 18 KIF5A disease-causing variants Therefore, the functional analysis of KIF1A variants on the located in the motor domain showed changes of the folding motility of the protein is crucial to understanding the effect of the free energy (11G) of the R169K variant in KIF1A, indicating variant on the phenotype. Even variants affecting the same amino that it may decrease protein stability (11G = 0.948) (Mahase acid can cause different phenotypes. Variants G102S, R167H, and et al., 2020). However, the R169T substitution was not tested. Frontiers in Neuroscience | www.frontiersin.org 7 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T The expression and purification of the motor domain constructs characterization of the molecular effect of the R169T variant in E. coli used in our study worked at fairly similar efficiencies, on the KIF1A protein together with the presence of the typical indicating that the R169T mutation may not cause a global clinical features indicates its causal pathogenic effect. misfolding of the motor domain. To assess the impact of the R169T variant on KIF1A functionality, we expressed and purified the motor domain (3– DATA AVAILABILITY STATEMENT 362 aa) as a 6× His fusion protein from E. coli. We have particularly chosen this minimal, monomeric construct, as it is The datasets presented in this study can be found in online known that regions following the motor domain are involved repositories. The names of the repository/repositories and in the auto-regulation of the motor activity and the dimeric accession number(s) can be found below: ENA, PRJEB41088. form is kept inactive by auto-inhibitory mechanisms. Using an ATPase assay, we could demonstrate that the R169T-mutated motor domain has an impaired ATPase activity compared with ETHICS STATEMENT the WT motor domain (Figure 2D). Consequently, the Kcat value is about 8-fold decreased in the R169T mutant. Furthermore, The studies involving human participants were reviewed and the strong increase in the Km value for microtubules suggests approved by the Institutional Ethics Committee of Institut that the impaired activity of the mutant motor domain R169T d’Investigació i Innovació Parc Taulí I3PT (CEIC 2019/514). is caused by a reduced stimulation of the ATPase activity by Written informed consent to participate in this study was microtubules. In support of this hypothesis, the 3D modeling of provided by the participants’ legal guardian/next of kin. Written the R169T variant predicts a disruption in the binding of KIF1A informed consent was obtained from the individual(s), and to β-tubulin, thereby suggesting that the lack of ATPase activity minor(s)’ legal guardian/next of kin, for the publication of any and motility is probably a consequence of weakened binding of potentially identifiable images or data included in this article. R169T to microtubules. Other missense variants reported in the motor domain and associated with NESCAV syndrome (T99M, R216C, and E253K) AUTHOR CONTRIBUTIONS show no motility in gliding assays, and in the case of the variant CA, SH, MM, MS, AS, and AR: design of the work. CA, SH, E253K, a dominant negative effect has been proven (Esmaeeli MM, AJ, EG, MG, MS, AS, and AR: data collection. CA, SH, Nieh et al., 2015). Dimeric kinesins like KIF1A require the activity AJ, MM, MS, AS, and AR: data analysis and interpretation. CA, of both motor domains for the processive movement along SH, AJ, MS, AS, and AR: manuscript preparation. All authors are microtubules. It is therefore tempting to speculate that like in the contributing to manuscript revision. case of E235K, the heterozygous R169T variant, which completely inhibits the ATPase activity of KIF1A, might exert a dominant negative effect. Several variants have been reported to be located in regions FUNDING involved in microtubule binding (loop 8 and loop 12). Among This work was supported by the Instituto de Salud Carlos III them, the de novo missense variant R167C is located within the (Grant No. PI16/01411 to MG and Grant No. CPII18/00027 same microtubule binding region 1 (β5-loop 8) as the R169T to AS), the Asociación Española de Síndrome de Angelman, variant reported here. 3D modeling of the R167C missense Institut d’investigació i innovació Parc Taulí I3PT (CIR2018/021), variant shows that it weakens the binding to microtubules, the Ministerio de Educación, Cultura y Deporte (Grant No. reducing the velocity and processivity of the kinesin (Lee et al., FPU16/01099 to MM), the Asociación Española Contra el Cáncer 2015; Scarabelli et al., 2015). Another variant, R307Q, located (Grant No. LABAE18009SEGU to MS), and the Centro de in the microtubule binding region 3 [α4-loop 12 (K-loop)-α5] Investigación Biomédica en Red Cancer - CIBERONC (Grant reduces the positive charge of the K-loop decreasing the affinity No. CB16/12/00363 to SH). Finally, AJ was supported by the of the motor domain to the microtubules (Hotchkiss et al., Wellcome Trust though a Senior Research Fellowship (202811). 2016). Recently, Kaur et al. reported a de novo missense variant (P305L) located in loop 12 of the motor domain and showed that the protein with the P305L variant accumulates in the cell ACKNOWLEDGMENTS body and that binding to microtubules was severely impaired (Kaur et al., 2020). We thank the patient’s family for their participation in this study In conclusion, we describe here the effect on the KIF1A motor and Thomas U. Mayer for reagents. domain of a novel de novo missense variant associated with severe developmental delay, spastic paraparesis, motor sensory neuropathy, bilateral optic nerve atrophy, progressive cerebellar SUPPLEMENTARY MATERIAL atrophy, epilepsy, ataxia, and hypotonia. We show that the R169T variant present in our patient strongly reduces ATPase The Supplementary Material for this article can be found online activity and therefore protein motility, probably due to defects at: https://www.frontiersin.org/articles/10.3389/fnins.2021. in microtubule binding as shown in structural modeling. The 618098/full#supplementary-material Frontiers in Neuroscience | www.frontiersin.org 8 May 2021 | Volume 15 | Article 618098 Aguilera et al. Novel KIF1A Missense Variant R169T REFERENCES Li, H., and Durbin, R. (2010). Fast and accurate long-read alignment with Burrows- Wheeler transform. Bioinformatics 26, 589–595. doi: 10.1093/bioinformatics/ Adzhubei, I. A., Schmidt, S., Peshkin, L., Ramensky, V. E., Gerasimova, A., Bork, P., btp698 et al. (2010). A method and server for predicting damaging missense mutations. Mahase, V., Sobitan, A., Johnson, C., Cooper, F., Xie, Y., Li, L., et al. (2020). Nat. Methods 7, 248–249. doi: 10.1038/nmeth0410-248 Computational analysis of hereditary spastic paraplegia mutations in the Chiba, K., Takahashi, H., Chen, M., Obinata, H., Arai, S., Hashimoto, K., kinesin motor domains of KIF1A and KIF5A. J. Theor. Comput. Chem. et al. (2019). Disease-associated mutations hyperactivate KIF1A motility and 19:2041003. doi: 10.1142/S0219633620410035 anterograde axonal transport of synaptic vesicle precursors. Proc. Natl. Acad. Nemani, T., Steel, D., Kaliakatsos, M., DeVile, C., Ververi, A., Scott, R., et al. Sci. U.S.A. 116, 18429–18434. doi: 10.1073/pnas.1905690116 (2020). KIF1A-related disorders in children: a wide spectrum of central and Choi, Y., and Chan, A. P. (2015). PROVEAN web server: a tool to predict the peripheral nervous system involvement. J. Peripher. Nerv. Syst. 25, 117–124. functional effect of amino acid substitutions and indels. Bioinformatics 31, doi: 10.1111/jns.12368 2745–2747. doi: 10.1093/bioinformatics/btv195 Okada, Y., Yamazaki, H., Sekine-Aizawa, Y., and Hirokawa, N. (1995). The neuron- DePristo, M. A., Banks, E., Poplin, R., Garimella, K. V., Maguire, J. R., Hartl, C., specific kinesin superfamily protein KIF1A is a unique monomeric motor for et al. (2011). A framework for variation discovery and genotyping using next- anterograde axonal transport of synaptic vesicle precursors. Cell 81, 769–780. generation DNA sequencing data. Nat. Genet. 43, 491–498. doi: 10.1038/ng. doi: 10.1016/0092-8674(95)90538-3 806 Pejaver, V., Urresti, J., Lugo-Martinez, J., Pagel, K. A., Lin, G. N., Nam, H.-J., et al. Esmaeeli Nieh, S., Madou, M. R. Z., Sirajuddin, M., Fregeau, B., Mcknight, D., Lexa, (2017). MutPred2: inferring the molecular and phenotypic impact of amino acid K., et al. (2015). De novo mutations in KIF1A cause progressive encephalopathy variants. bioRxiv [Preprint]. doi: 10.1101/134981 bioRxiv: 134981, and brain atrophy. Ann. Clin. Transl. Neurol. 2, 623–635. doi: 10.1002/acn3.198 Pennings, M., Schouten, M. I., van Gaalen, J., Meijer, R. P. P., de Bot, S. T., Hammond, J. W., Cai, D., Blasius, T. L., Li, Z., Jiang, Y., Jih, G. T., et al. (2009). Kriek, M., et al. (2020). KIF1A variants are a frequent cause of autosomal Mammalian Kinesin-3 motors are dimeric in vivo and move by processive dominant hereditary spastic paraplegia. Eur. J. Hum. Genet. 28, 40–49. doi: motility upon release of autoinhibition. PLoS Biol. 7:e72. doi: 10.1371/journal. 10.1038/s41431-019-0497-z pbio.1000072 Rentzsch, P., Witten, D., Cooper, G. M., Shendure, J., and Kircher, M. (2019). Hirokawa, N., Pfister, K. K., Yorifuji, H., Wagner, M. C., Brady, S. T., and Bloom, CADD: predicting the deleteriousness of variants throughout the human G. S. (1989). Submolecular domains of bovine brain kinesin identified by genome. Nucleic Acids Res. 47, D886–D894. doi: 10.1093/nar/gky1016 electron microscopy and monoclonal antibody decoration. Cell 56, 867–878. Reva, B., Antipin, Y., and Sander, C. (2007). Determinants of protein function doi: 10.1016/0092-8674(89)90691-0 revealed by combinatorial entropy optimization. Genome Biol. 8:R232. doi: Hirokawa, N., and Tanaka, Y. (2015). Kinesin superfamily proteins (KIFs): various 10.1186/gb-2007-8-11-r232 functions and their relevance for important phenomena in life and diseases. Richards, S., Aziz, N., Bale, S., Bick, D., Das, S., Gastier-Foster, J., et al. (2015). Exp. Cell Res. 334, 16–25. doi: 10.1016/j.yexcr.2015.02.016 Standards and guidelines for the interpretation of sequence variants: a joint Hotchkiss, L., Donkervoort, S., Leach, M. E., Mohassel, P., Bharucha-Goebel, consensus recommendation of the American College of Medical Genetics D. X., Bradley, N., et al. (2016). Novel De Novo mutations in KIF1A as and Genomics and the Association for Molecular Pathology. Genet. Med. 17, a cause of hereditary spastic paraplegia with progressive central nervous 405–423. doi: 10.1038/gim.2015.30 system involvement. J. Child Neurol. 31, 1114–1119. doi: 10.1177/0883073816 Scarabelli, G., Soppina, V., Yao, X.-Q., Atherton, J., Moores, C. A., Verhey, 639718 K. J., et al. (2015). Mapping the processivity determinants of the kinesin- Huo, L., Yue, Y., Ren, J., Yu, J., Liu, J., Yu, Y., et al. (2012). The CC1-FHA tandem 3 motor domain. Biophys. J. 109, 1537–1540. doi: 10.1016/j.bpj.2015. as a central hub for controlling the dimerization and activation of kinesin-3 08.027 KIF1A. Structure 20, 1550–1561. doi: 10.1016/j.str.2012.07.002 Sundaram, L., Gao, H., Padigepati, S. R., McRae, J. F., Li, Y., Kosmicki, J. A., et al. Ioannidis, N. M., Rothstein, J. H., Pejaver, V., Middha, S., McDonnell, S. K., Baheti, (2018). Predicting the clinical impact of human mutation with deep neural S., et al. (2016). REVEL: an ensemble method for predicting the pathogenicity networks. Nat. Genet. 50, 1161–1170. doi: 10.1038/s41588-018-0167-z of rare missense variants. Am. J. Hum. Genet. 99, 877–885. doi: 10.1016/j.ajhg. Wang, K., Li, M., and Hakonarson, H. (2010). ANNOVAR: functional annotation 2016.08.016 of genetic variants from high-throughput sequencing data. Nucleic Acids Res. Karczewski, K. J., Francioli, L. C., Tiao, G., Cummings, B. B., Alföldi, J., Wang, Q., 38:e164. doi: 10.1093/nar/gkq603 et al. (2020). The mutational constraint spectrum quantified from variation in Wang, Z., and Moult, J. (2001). SNPs, protein structure, and disease. Hum. Mutat. 141,456 humans. Nature 581, 434–443. doi: 10.1038/s41586-020-2308-7 17, 263–270. doi: 10.1002/humu.22 Kaur, S., Van Bergen, N. J., Verhey, K. J., Nowell, C. J., Budaitis, B., Yue, Y., Woehlke, G., Ruby, A. K., Hart, C. L., Ly, B., Hom-Booher, N., and Vale, R. D. et al. (2020). Expansion of the phenotypic spectrum of de novo missense (1997). Microtubule interaction site of the kinesin motor. Cell 90, 207–216. variants in kinesin family member 1A (KIF1A). Hum. Mutat. 41, 1761–1774. doi: 10.1016/s0092-8674(00)80329-3 doi: 10.1002/humu.24079 Zhang, Z., Miteva, M. A., Wang, L., and Alexov, E. (2012). Analyzing effects Kikkawa, M., Okada, Y., and Hirokawa, N. (2000). 15 A resolution model of of naturally occurring missense mutations. Comput. Math. Methods Med. the monomeric kinesin motor. KIF1A. Cell 100, 241–252. doi: 10.1016/s0092- 2012:805827. doi: 10.1155/2012/805827 8674(00)81562-7 Klopfenstein, D. R., Tomishige, M., Stuurman, N., and Vale, R. D. (2002). Role Conflict of Interest: The authors declare that the research was conducted in the of phosphatidylinositol(4,5)bisphosphate organization in membrane transport absence of any commercial or financial relationships that could be construed as a by the Unc104 kinesin motor. Cell 109, 347–358. doi: 10.1016/s0092-8674(02) potential conflict of interest. 00708-0 Lee, J.-R., Shin, H., Ko, J., Choi, J., Lee, H., and Kim, E. (2003). Characterization of Copyright © 2021 Aguilera, Hümmer, Masanas, Gabau, Guitart, Jeyaprakash, the movement of the kinesin motor KIF1A in living cultured neurons. J. Biol. Segura, Santamaria and Ruiz. This is an open-access article distributed under the Chem. 278, 2624–2629. doi: 10.1074/jbc.M211152200 terms of the Creative Commons Attribution License (CC BY). The use, distribution Lee, J. R., Srour, M., Kim, D., Hamdan, F. F., Lim, S. H., Brunel-Guitton, C., et al. or reproduction in other forums is permitted, provided the original author(s) and (2015). De novo mutations in the motor domain of KIF1A cause cognitive the copyright owner(s) are credited and that the original publication in this journal impairment, spastic paraparesis, axonal neuropathy, and cerebellar atrophy. is cited, in accordance with accepted academic practice. No use, distribution or Hum. Mutat. 36, 69–78. doi: 10.1002/humu.22709 reproduction is permitted which does not comply with these terms. Frontiers in Neuroscience | www.frontiersin.org 9 May 2021 | Volume 15 | Article 618098