Structural and functional insights of the human peroxisomal ABC transporter ALDP

  1. Yutian Jia
  2. Yanming Zhang
  3. Wenhao Wang
  4. Jianlin Lei
  5. Zhengxin Ying  Is a corresponding author
  6. Guanghui Yang  Is a corresponding author
  1. State Key Laboratory for Agrobiotechnology, Department of Nutrition and Health, College of Biological Sciences, China Agricultural University, China
  2. Technology Center for Protein Sciences, Ministry of Education Key Laboratory of Protein Sciences, School of Life Sciences, Tsinghua University, China

Abstract

Adrenoleukodystrophy protein (ALDP) is responsible for the transport of very-long-chain fatty acids (VLCFAs) and corresponding CoA-esters across the peroxisomal membrane. Dysfunction of ALDP leads to peroxisomal metabolic disorder exemplified by X-linked adrenoleukodystrophy (ALD). Hundreds of ALD-causing mutations have been identified on ALDP. However, the pathogenic mechanisms of these mutations are restricted to clinical description due to limited structural and biochemical characterization. Here we report the cryo-electron microscopy structure of human ALDP with nominal resolution at 3.4 Å. ALDP exhibits a cytosolic-facing conformation. Compared to other lipid ATP-binding cassette transporters, ALDP has two substrate binding cavities formed by the transmembrane domains. Such structural organization may be suitable for the coordination of VLCFAs. Based on the structure, we performed integrative analysis of the cellular trafficking, protein thermostability, ATP hydrolysis, and the transport activity of representative mutations. These results provide a framework for understanding the working mechanism of ALDP and pathogenic roles of disease-associated mutations.

Editor's evaluation

The adrenoleukodystrophy protein (ALDP) or ABCD1 is an ABC transporter that participates in the transport of free very long-chain fatty acids and their CoA esters across the peroxisomal membrane. By determining the cryo EM structure of human ABCD1 the study represents a valuable insight into its transport mechanism and the mechanistic basis for mutations causing the severe neurodegenerative disorder, X-linked adrenoleukodystrophy. The structure and functional studies of disease-causing mutations are solid and will appeal to the transporter and medical genetics communities.

https://doi.org/10.7554/eLife.75039.sa0

Introduction

X-linked adrenoleukodystrophy (ALD) is an inherited disease characterized by progressive demyelination of the central nervous system (Berger et al., 2014). Loss of myelin slows down the transmission of nerve impulses and triggers neuroinflammation (Ferrer et al., 2010; Kettwig et al., 2021; Singh et al., 2009). Pathogenesis of ALD is tightly associated with mutations on the adrenoleukodystrophy protein (ALDP), which transports very-long-chain fatty acid-CoAs (VLCFA-CoAs) from cytosol into the peroxisome (Figure 1; Mosser et al., 1993). Over 900 disease-derived mutations have been identified on ALDP (https://adrenoleukodystrophy.info/mutations-biochemistry/mutations-biochemistry). ALDP, also known as ABCD1, belongs to the ATP-binding cassette sub-family D. The other three members are ALDP-related protein (ALDRP/ABCD2; Lombard-Platet et al., 1996), PMP70/ABCD3 (Kamijo et al., 1990), and a cobalamin transporter ABCD4 (Coelho et al., 2012). ABCD1–3 are distributed on peroxisomes, while ABCD4 is localized on lysosomes (Xu et al., 2019). ALDP and ALDRP transport VLCFA with different specificities (van Roermund et al., 2011). ALDRP may compensate for the function loss of ALDP (Holzinger et al., 1997; Liu et al., 1999; Lombard-Platet et al., 1996). PMP70 was reported to transport LCFA-CoA (Ranea-Robles et al., 2021; van Roermund et al., 2014).

Figure 1 with 4 supplements see all
Functional and structural characterization of human adrenoleukodystrophy protein (ALDP).

(A) Schematic of very-long-chain fatty acids transport into the peroxisomes by ALDP. (B–C) ATP hydrolysis of ALDP stimulated by C22:0-CoA (data are represented as mean ± SD; n=3; three biological repeats). (D) Cryo-electron microscopy structure of ALDP. The transmembrane domains (TMDs) and nucleotide-binding domains (NBDs) are indicated.

Figure 1—source data 1

Cryo-electron microscopy data collection, refinement, and validation statistics.

https://cdn.elifesciences.org/articles/75039/elife-75039-fig1-data1-v2.doc

Structure-based functional characterization of ALDP is eagerly required for understanding the VLCFA transport mechanism and the pathogenic roles of ALD-derived mutations. To address these questions, we first performed ATP hydrolysis assay using purified ALDP. In the presence of C22:0-CoA, ALDP exhibits robust ATP hydrolysis with the Vmax of ~193 ± 8.2 mol Pi min−1 mol−1 protein and Km value of ~0.17 µM C22:0-CoA (Figure 1B and C). The structure of human ALDP reveals an assembly of two identical subunits that exhibit a domain-swapped arrangement (Figure 1D and Figure 1—figure supplements 12 and Figure 1—source data 1). Each subunit contains six transmembrane helices (TMs; Figure 1D and Figure 1—figure supplement 2). A short helix located at the peroxisomal side (extracellular helix [EH]) is a featured structural element of ALDP based on sequence alignment (Figure 1—figure supplement 4). Two coiled-coil domain-like densities at the C-terminus were observed (C-terminal helix [CH]). The local resolution of these densities is estimated to be ~4.0 Å, which is not accurate to build an atomic model.

During submission of this study, four other groups independently reported ALDP structures in different states (Chen et al., 2022; Le et al., 2022; Wang et al., 2022; Li et al., 2021). Our structure is the ligand-free state as no substrates were supplemented during purification. We observed two unknown densities at the crevice of TMDs, which were also reported in the ligand-free structure by Chen et al., 2022. When superimposed to the oleoyl-CoA- or C22:0-CoA-bound structures, the unknown densities partially overlap with the substrates (Figure 1—figure supplement 3). Such observation indicates a probable extrusion of the unknown densities during substrate loading. Structure comparison of our structure with the C22:0-CoA-bound and ATP-bound structure reveals dramatic conformational changes during the transport cycle. For instance, the distance between TM4 and TM6 at the juxta-membrane region changes from 23.6 Å to 13.1 Å upon substrate binding and further gets closed to 5.7 Å in the presence of ATP. The nucleotide-binding domains (NBDs) undergo continuously spatial rearrangement. Consequently, the TM4-6 and TM1-2 bend to open the peroxisomal side exit for substrate (Figure 2). Conformational change of TMDs is accompanied with the movement of NBDs and the disappearance of EH or CH.

Conformational changes of adrenoleukodystrophy protein (ALDP) during the whole transport cycle.

(A) Structural alignments of ALDP in ligand-free, C22:0-CoA-bound (7VZB) (Chen et al., 2022), and ATP-bound states (7SHM) (Wang et al., 2022). The distance change (in Å) between transmembrane helix 4 (TM4) and TM6 reflects the conformational changes during lipid transport. The hollow arrows indicate three views of ALDP that are enlarged in B–D. CH: C-terminal helix; EH: extracellular helix.

Structure determination of ALDP allows classification of ALD-associated mutations. 970 ALD-associated mutations in ALDP affecting 232 residues have been reported. Among these 232 residues, 88 harbor mutations to two or more types of amino acids, which we defined as hotspot residues. Based on the structural mapping, the 88 hotspot residues can be roughly classified into 3 groups (Figure 3B). The first group of 10 residues lines along the substrate-binding cavity. These residues likely play a role in substrate coordination and delivery. The second group of 35 residues is located in other region of the TMDs, which undergoes conformational changes. The third group of 43 residues is located on the NBDs that may have influence on ATP hydrolysis (Figure 3B and Figure 3—figure supplement 1).

Figure 3 with 5 supplements see all
Functional analysis of adrenoleukodystrophy (ALD)-associated mutations.

(A) Conformational changes of adrenoleukodystrophy protein (ALDP) from ligand-free to ATP-bound state. (B) Structural mapping of clinical-derived pathogenic mutations of ALDP. The Cα atoms of the indicated residues are shown as spheres. The residues involved in substrate binding are shown as red spheres. Other mutation sites on transmembrane domains are presented in green. Mutations on nucleotide-binding domain (NBD) and the single-mutation T693M on C-terminal helix are colored cyan, respectively. (C) Spatial location of selected ALD mutants and rationally designed mutations. (D) Transport of cytosolic C22:0-CoA, C24:0-CoA, C26:0-CoA, and ATP hydrolysis by wild-type (WT) ALDP and mutations (data are represented as mean ± SD; n = 3; three biological repeats). EH: extracellular helix.

To examine how ALD-associated mutations affect the function of ALDP, we selected those reported to have normal expressed levels but dysfunction for integrative analysis (Coll et al., 2005; Feigenbaum et al., 1996; Guimarães et al., 2002). These mutations include W339R, S342P, G343V, A396T, Q544R, and T693M (Imamura et al., 1997; Kemp et al., 2001; Lan et al., 2011; Liu et al., 2022; Pan et al., 2005; Takahashi et al., 2007; Watkins et al., 1995; Wichers et al., 1999). In addition, to evaluate the role of different structure elements, we designed four constructs based on structural changes. The S164/Y310 residues on TMDs, anchored by a hydrogen bond, both move quite large distance during the transport cycle but remain relatively static distance between themselves. The E291/R518 residues are mapped to the interface between TMD and NBD. The side chain of E291 stretches into the NBD of another protomer, while R518 binds to the α-phosphate of ATP. These two pairs of residues represent two kinds of dimeric interfaces. We mutated these residues into alanine to examine how disturbance of the dimeric assembly affects the activity. Another two constructs are truncation of EH (Δ364–374) or CH (Δ683–745).

We first measured the VLCFA-CoA transport. Different mutations are individually expressed in HEK293 cells. The cytosolic C22:0-CoA, C24:0-CoA, and C26:0-CoA were measured by quantitative mass spectrometry (Wang et al., 2019), and the transport ability was calculated (Figure 3C). Because the substrate transport is coupled with ATP hydrolysis, we also measured the ATPase activity of these mutants using purified proteins (Figure 3C). To exclude the possibility that the mutants may alter the cellular localization and thermostability compared to wild-type (WT), we investigated their subcellular localization and measured the Tm values of purified proteins. Compared with A95D, a disease-causing mutation known to not colocalize with peroxisome and reduce the C24:0 β-oxidation (Morita et al., 2013), all the selected mutants show puncta distribution and colocalize with catalase, suggesting their normal trafficking to the peroxisomes (Figure 3—figure supplement 2). The similar Tm values of purified mutant proteins rule out the instability-caused functional abrogation (Figure 3—figure supplement 3).

The transport activity and the ATP hydrolysis of the mutants are normalized to WT. All six disease-derived mutants exhibit decreased transport toward three species of VLCFA-CoA. Five mutants show abrogated ATP hydrolysis except for the mutation S342P. The ALD mutation T693M is mapped on the CH. Because the CH cannot be observed in the ATP-bound state, to evaluate the potential pathogenic role of the T693M mutant, we generated the homomeric model of ALDP by AlphaFold (hereafter as AF-model) (Evans et al., 2022; Jumper et al., 2021). The predicted model of ALDP displays similar conformation to the ATP-bound state. Albeit the part of CH in AF-model needs further validation, the prediction may provide potential clues for interpreting the T693M mutation at the current stage. In the AF-model, T693 is located near W664, which is surrounded by ALD mutants on the NBD. The distance between the side chains of T693 and W664 can reach in 3.4 Å (Figure 3—figure supplement 4). Replacement of Thr by Met residue with large side chain may affect the neighboring residues near W664 to disturb the ATP hydrolysis. For the four designed constructs, alanine substitution of S164/Y310 and E291/R518 has negligible influence on the ATP hydrolysis. By contrast, the lipid transport decreased because destabilization of the interaction between these residues may disturb the conformational changes during substrate translocation. As one featured structural element, deletion of the EH (Δ364–374) may generate structural constraints between TM5 and TM6 during the conformational change of ALDP (Chen et al., 2022; Le et al., 2022; Wang et al., 2022). Truncation of CH (Δ683–745) does not change the assembly of NBD but affects the conformational changes between the two protomers, thus the ATP hydrolysis level slightly decreased with abrogated lipid transport.

To gain insights into the transport of different lipids by ABC transporters, we compared ALDP with the phospholipid transporter ABCB4 (Nosol et al., 2021; Olsen et al., 2020). Typically, ALDP harbors two substrate-binding pockets that expand perpendicularly to the membrane, but ABCB4 binds one phosphatidylcholine (PC) (Figure 3—figure supplement 5). The dimension of one binding pocket is ~1695 Å3 for C22:0-CoA and ~1236 Å3 for PC calculated by ProteinPlus (Schöning-Stierand et al., 2020). The different formulas between VLCFA-CoA and PC may provide clues for the different cavities. Regardless of the flexibility, the length of C22:0-CoA (~55 Å) is larger than PC (~40 Å) (Figure 3—figure supplement 5B). The PC molecule has two acyl tails that can be coordinated by ABCB4. The VLCFAs have only one hydrophobic tail. More residues in a larger binding pocket of ALDP may help in stabilizing the flexible long tail of VLCFA-CoA. The EH and CH facilitate conformational changes to expand the binding pocket. Taken together, the integrative characterization of ALDP provides a framework to illustrate the molecular basis of VLCFA transport and the function of ALD-derived mutations.

Methods

Protein expression and purification

The coding sequence of human ALDP was subcloned into the pFastBac1 vector with a N-terminal Flag tag. The recombinant ALDP was expressed using the Bac-to-bac system (Invitrogen). Briefly, bacmids were generated in DH10Bac competent cells. The resulting baculoviruses were amplified in Sf9 insect cells (Invitrogen). ALDP was overexpressed in Sf9 insect cells grown in the serum-free medium (Gibco). 60 hr after P3 virus infection, transfected cell pellet was collected and homogenized in the buffer containing 25 mM Tris-HCl, pH 7.4, 150 mM NaCl supplemented with 1 mM phenylmethylsulfonyl fluoride, 1.3 μg ml−1 aprotinin, 0.7 μg ml−1 pepstatin, and 5 μg ml−1 leupeptin. After brief sonication, the suspension was supplemented with 2,2-didecylpropane-1,3-bis-β-D-maltopyranoside (LMNG, Anatrace) to a final concentration of 1% (w/v) and cholesteryl hemisuccinate Tris salt (CHS, Anatrace) to 0.1% (w/v). After incubation at 4°C for 2 hr, the mixture was centrifuged at 150,000 g for 30 min. The supernatant was mixed with anti-Flag M2 affinity gel (Sigma). The gel was rinsed with 25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.005% LMNG, and 0.0005% CHS. The target protein was eluted with wash buffer plus 200 μg ml−1 flag peptide and further purified through gel-filtration chromatography in the buffer containing 25 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 0.06% digitonin. Peak fractions were concentrated for cryo-EM analysis.

Preparation of the cryo-EM samples

Aliquots of 4 µl ALDP was dropped onto glow-discharged holey carbon grids (Quantifoil Au R1.2/1.3, 300 mesh). The grids were blotted for 3 s and flash-frozen in liquid ethane using Vitrobot Mark IV (FEI). The sample was imaged on an FEI 300 kV Titan Krios transmission electron microscope equipped with a Cs corrector and Gatan GIF Quantum energy filter (slit width 20 eV), recorded by a Gatan K2 Summit detector with a nominal magnification of ×640,000. A series of defocus values from –1.5 to –1.8 µm was used during data collection. Each image was dose-fractionated to 32 frames with a total electron dose of ~50 e Å−2 and a total exposure time of 5.6 s. AutoEMation II (developed by J. Lei) (Lei and Frank, 2005) was used for fully automated data collection. All stacks were motion-corrected using MotionCor2 with a binning factor of 2 (Zheng et al., 2017), resulting in a pixel size of 1.0979 Å. The defocus values were estimated using Gctf (Zhang, 2016), and dose weighing was performed concurrently (Grant and Grigorieff, 2015).

Cryo-EM data processing and model building

In total, 1,259,972 particles were auto-picked from these 1922 movie stacks using Gautomatch (developed by Kai Zhang, http://www.mrc-lmb.cam.ac.uk/kzhang/Gautomatch) (Figure 1—figure supplement 1). After two-dimensional classification, 472,168 particles were selected to generate the initial model with a mask diameter of 200 Å in C1 symmetry. To avoid model bias, the initial model was low-pass filtered for the following three-dimensional (3D) classification. The 472,168 particles were subjected to 50 iterations of global angular search 3D classification. Each of the 50 iterations has one class and a step size of 7.5°. For each of the last five iterations (iterations 46–50) of the global search, the local angular search 3D classification was executed with a class number of 3, a step size of 3.75°, and a local search range of 15°. A total of 289,321 particles were selected from the local angular searching 3D classification after removing the redundant particles and particles from bad classes (Figure 1—figure supplement 1). The selected particles were subjected to 3D autorefinement, yielding a density map with average resolution at 3.6 Å. After multi-reference 3D classification, 135,662 particles were selected for autorefinement with C2 symmetry, resulting a map at 3.4 Å resolution. The structure of ABCD4 was used as a template for the model building of ALDP. The structure was refined in real space using PHENIX with secondary structure and geometry restraints (Adams et al., 2002). The atomic model was manually improved using COOT (Emsley and Cowtan, 2004).

Determination of cytosolic very-long-chain fatty acids

Sample preparation and liquid chromatography-mass spectrometry (LC-MS/MS) conditions as described (Wang et al., 2019). Chromatographic separation was achieved using Q Exactive HF LC-MS/MS system and Phenomenex Luna 5 µm C5 column (i.d. 100×2.0 mm). Mobile phase A contained HPLC-grade H2O-ACN 40/60 (v/v) with 10 mM ammonium acetate, and mobile B was isopropanol-ACN 90/10 (v/v). Both WT ALDP and mutant constructs were subcloned into a pCAG vector and expressed in HEK293F cells (Invitrogen). Transfected cells were harvested after 60 hr. 2.0×107 cells were collected and washed by PBS. 1 mL ACN and 10 μL internal standard mix solution were added to samples or quality control sample (expressed by empty vector). After sonication and centrifugation, 50 μL hydrochloric acid (12 mol/L) was added into the supernatant then VLCFA was hydrolyzed at 70°C for 1 hr. Each sample was extracted with 2 ml hexane, the dried residue was reconstituted in 100 μL methanol. 5 μL aliquots of the reconstitutes were loaded onto the LC–MS/MS system for analysis.

Immunofluorescence cytochemistry

Hela cells fixed in 4% Paraformaldehyde (20 min) were permeabilized with 0.1% Triton X-100 in PBS (30 min, room temperature), blocked with SuperBlock (30 min), before incubation with rabbit anti-catalase antibody (1:500; Abcam, ab16731). Immunoreactivity was visualized by AlexaFluor-488-conjugated goat anti-rabbit IgG (1:1000; 1 hr at room temperature).

ATPase activity assay

For determining ATP hydrolysis capacity, wtABCD1 and mutants were deleted N-terminal 54 residues to improve stability. The malachite green ATPase assay was chosen to exhibit the ABCD1 ATPase activity in the presence of C22:0-CoA (behenoyl coenzyme A, ammonium salt) as previously described (Baykov et al., 1988). Purified ABCD1 in LMNG/CHS micelles was pre-incubated on ice for 10 min in 20 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM DTT, 2 mM MgCl2 with different concentrations of C22:0-CoA. The 100 µl reaction started by supplementing with 2 mM ATP and then carried out at 37°C for 30 min, after that terminated by adding 25 µl fresh Gold-mix solution (1 mM malachite green, 1.2% ammonium molybdate, and 0.15% Tween 20). The mixture was incubated at RT for 30 min before being detected absorbance in 96-well micro-plate at 630 nm. Statistical analysis was performed using GraphPad Prism 7.

The thermal stability assay

The thermal stability assay was performed by measuring intrinsic tryptophan fluorescence and light back-scattering for fragment screening by a Prometheus NT.48 device (NanoTemper Technologies GmbH, Munich, Germany) (Ahmad et al., 2021). wtABCD1 and mutations were purified in LMNG/CHS and concentrated to 0.5 mg/ml, after centrifuged at 14,000 × g for 15 min at 4°C, supernatant was loaded with standard capillaries (NanoTemper Technologies GmbH, Munich, Germany; Cat# PR-C002) into the Prometheus device and subjected to a linear thermal ramp (1°C/min, from 25 to 95°C) and collected fluorescence at 350 nm and 330 nm. Unfolding transition midpoints were determined automatically from the first derivative of the fluorescence ratio (F350/F330). Data analysis was performed using PR.Therm Control (NanoTemper Technologies GmbH) and GraphPad Prism 7.

Data availability

Cryo-EM data have been deposited in PDB under the accession code 7VR1. All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for Figure 1 and Figure 3—figure supplement 3.

The following data sets were generated
    1. Yang GH
    2. Jia YT
    3. Zhang YM
    (2022) RCSB Protein Data Bank
    ID 7VR1. Cryo-EM structure of the ATP-binding cassette sub-family D member 1 from Homo sapiens.
The following previously published data sets were used

References

    1. Emsley P
    2. Cowtan K
    (2004) Coot: model-building tools for molecular graphics
    Acta Crystallographica. Section D, Biological Crystallography 60:2126–2132.
    https://doi.org/10.1107/S0907444904019158
    1. Feigenbaum V
    2. Lombard-Platet G
    3. Guidoux S
    4. Sarde CO
    5. Mandel JL
    6. Aubourg P
    (1996)
    Mutational and protein analysis of patients and heterozygous women with X-linked adrenoleukodystrophy
    American Journal of Human Genetics 58:1135–1144.
    1. Watkins PA
    2. Gould SJ
    3. Smith MA
    4. Braiterman LT
    5. Wei HM
    6. Kok F
    7. Moser AB
    8. Moser HW
    9. Smith KD
    (1995)
    Altered expression of ALDP in X-linked adrenoleukodystrophy
    American Journal of Human Genetics 57:292–301.

Decision letter

  1. David Drew
    Reviewing Editor; Stockholm University, Sweden
  2. Kenton J Swartz
    Senior Editor; National Institute of Neurological Disorders and Stroke, National Institutes of Health, United States
  3. David Drew
    Reviewer; Stockholm University, Sweden
  4. Konstantinos Beis
    Reviewer

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Structure insights of the human peroxisomal ABC transporter ALDP" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including David Drew as Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Kenton Swartz as the Senior Editor. The following individual involved in review of your submission have agreed to reveal their identity: Konstantinos Beis (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1. The modelled two putative Co-A esters lack the positively charged residues for coordinating phosphates. Indeed, the Co-A binding in the human ABCD1 structure by Chen and co-workers make more chemical sense https://www.biorxiv.org/content/10.1101/2021.09.24.461565v2.full.pdf. This study found they same shaped density as you have seen in your ABCD1 structure, but have modelled the lipid 18:0 Lyso PE molecule into this location instead. Judging the deposited maps, we think that this is the more likely lipid for your density, rather than CoA as proposed. We need to see an update of the substrate-binding site and the corresponding mutations to either the new substrate-binding site position as in the competing paper or more solid experimental data for keeping with the currently modelled substrate-binding site.

2. Certain assumptions have been made but there is no references or experimental validation from this work to warrant such conclusions; eg p6 'Pathogenic mutations on these residues may disturb the stability of ALDP. The residues of the third group are located on the NBDs, which may hinder the hydrolysis of nucleotides. Given the explosion of structural information on ABC transporters (for example, of the ~50 human ABC transporters, structures have been reported for the human A1, A4, B1, B2, B3, B4, B6, B8, B10, B11, C7, D4, G2, G5, G8 transporters in addition to D1), it is a challenge to interpret new results in the context of existing knowledge about ABC transporters. The transport assays provide an approach to connecting structure and function, but too little information is provided about the assays to make those connections. Of the ~900 mutants, why were the variants in figure 4B selected? Where are they located in the structure? What do the results of the transport assay mean? For example, is the reduction of C22:S-CoA in the cytosol for the G343V mutant due to impaired transport or to reduced expression, misfolding or some other cause? mutations are mainly mapped to the TMDs, which may affect the conformational changes during substrate transport.

Please show the correct trafficking, expression (normalised to WT) and folding of all mutations tested in HEK293 cells.

Please show that the purified protein is capable of turning over ATP and that the basal ATPase activity stimulated by the substrate. We would recommend carrying out complementary ATPase assays on key mutations also. It is critical to better integrate the mutagenesis data with the structure.

3. Please provide some rationale basis for the additional structural features found in ABCD1 (peroxisomal helix and the C-terminal coil-coiled domain) and how this might be related to its function?

4. The final model is difficult to follow and this, of course, relates to the "actual" location of the substate-binding site. Is an outward-facing conformation to be expected in ABCD1? Furthermore, the conformational changes as proposed by comparing to AlphaFold models should be made with caution until experimental validation of these states. Please incorporate the AlphaFold models more carefully into the paper, i.e.,what are the state-specific co-evolved residue-pairs predicted in the AlphaFold models, do these make sense, and can these contact pairs be experimentally validated by mutagenesis?

5. The paper needs to be extensively re-written. We would like to see how the structure uncovers new mechanistic insights. With the additional functional data requested, consider how the ABCD1 structure enables a understanding of why very long chain fatty acids use such an ABC transporter structures for being imported into peroxisomes. How does this differ to other ABC transporters that transport lipids and why?

Reviewer #1 (Recommendations for the authors):

1. The modelled two putative Co-A esters lack the positively charged residues for coordinating phosphates. Indeed, the Co-A binding in the human ABCD1 structure by Chen and co-workers make more chemical sense https://www.biorxiv.org/content/10.1101/2021.09.24.461565v2.full.pdf. This study found they same shaped density as you seen in your ABCD1 structure, but have modelled the lipid 18:0 Lyso PE molecule into this location instead. Judging the deposited maps I think that this is the more likely lipid for your density, rather than CoA as proposed. Please read this submission more carefully and update accordingly as I find it worrying that this has not been picked up.

2. Please check that ABCD1 mutations are trafficked correctly to peroxisomes. Also one needs to normalise uptake against expression level. Please complement substrate uptake experiments with ATPase activity measurements also.

3. This paper is difficult to follow in places and using AlphaFold structures to make conclusions regarding local conformational changes seems premature at this stage without further validation.

4. Please provide some rationale basis for the additional structural features found in ABCD1 (peroxisomal helix and the C-terminal coil-coiled domain) and how this might be related to its function?

Reviewer #3 (Recommendations for the authors):

The CoA-ester binding should be discussed in more details.

Although several mutants have been prepared, they are not really discussed in the context of the manuscript or previously published work. The discussion needs to expand further. Certain assumptions have been made but there is no references or experimental validation from this work to warrant such conclusions; eg p6 'Pathogenic mutations on these residues may disturb the stability of ALDP. The residues of the third group are located on the NBDs, which may hinder the hydrolysis of nucleotides. The other mutations are mainly mapped to the TMDs, which may affect the conformational changes during substrate transport.'. Based on their structure it should be possible to explain better why and how the mutants may affect the function.

The authors should show that the purified protein is capable of turning over ATP and that the basal ATPase activity stimulated by the substrate.

Figure 4B; are all mutants expressing at similar levels as the WT protein? They authors should show a Western Blot. Additionally, what do the error bars represent?

The authors should briefly state in the M and M how they prepared the Alphafold models; ie collar or full Alphafold. Why not use the Alphafold multimer and make the homodimer?

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Structural and functional insights of the human peroxisomal ABC transporter ALDP" for further consideration by eLife. Your revised article has been evaluated by Kenton Swartz (Senior Editor) and a Reviewing Editor.

The manuscript has been improved but there are some remaining issues that need to be addressed, as outlined below:

– The main impact of the paper is the ALDP structure with a number of supporting mutations. The current Manuscript word length is around 1800 words and 4 figures, which is close to the requirements for a short report (1500 words and 3 figures). We think the paper would come across much better as a short report, and ask that you please re-write to meet these requirements and improve the readability of the text in the process.

https://reviewer.elifesciences.org/author-guide/types

The referees have further concerns to be addressed:

– line 100: "The TM2 serves as a mechanical lever: the cytosolic side moves towards the central cavity, while the peroxisomal side may move backwards the symmetry axis"

Two points:

i. is TM2 really a mechanical level? A lever is a rigid object with a fulcrum point that amplifies an input force to become an output force. Is this what is meant? If so, where is the fulcrum and what are the input and output forces?

ii. I don't understand what is being described in the second part of this sentence.

– Line 148 "Intriguingly, there are four constructs exhibit unaffected or little enhanced ATPase hydrolysis but obstructed lipid transport"

– Line 164 " Truncation of the CH does not change the completeness of the NBD"

What is meant by the completeness of the NBD?

– Line 194 – what precisely do these dimensions (42 Å or 40 Å) describe? – this implies that the TMDs have similar dimensions to accommodate their substrates – but lines 203 (and the abstract) state that the binding pocket of ALDP is much larger than ABCB4. Please clarify.

What are the volumes of the binding cavities in ALDP and ABCB4, and how does the volume of PC compare to the volume of the VLCFA substrate?

– Line 264 – Figure 4 – which PDBs are which? (ie ligand-free, substrate-bound, ATP-bound)?

– One last point that likely reflects an issue on my end, but I was not able to see the EM map postrun9.mrc in coot 0.9.6 – Coot opened the map and provided a histogram of the density, but it didn't display anything over the "A_fit_real_space_refined-coot-34_real_space_refined_001_all_states.pdb" coordinates.

– Why is the structure in line 190 compared to ABCB4 and not the other structures listed in line 90? It makes more sense to discuss the mechanism based on the available structures of the same family in addition to the ABCB4.

– It is still unclear how they selected their mutants; lines 122-125 do not have any refs of previously identified mutations. Do we know any of them being specific disease related?

– Lines 85-91 contradict each other; they state that there is lipid-like density but at the same time they write that they have an apo structure.

Reviewer #2 (Recommendations for the authors):

The structure of human ALDP is of significance given the role of this transporter in the transport of VLCFAs across the peroxisomal membrane. Nearly a thousand ALD-causing mutations have been identified for ALDP and the structure provides a framework for assessing their functional consequences. In this revised manuscript, the authors have addressed many of the reviewers' concerns in the rebuttal letter. Given the importance of this transporter, particularly the human version of ALDP, I am sympathetic to the publication.

The manuscript, however, needs extensive polishing to be suitable for publication, however. Here are a few examples

line 100: "The TM2 serves as a mechanical lever: the cytosolic side moves towards the central cavity, while the peroxisomal side may move backwards the symmetry axis"

Two points:

i. is TM2 really a mechanical level? A lever is a rigid object with a fulcrum point that amplifies an input force to become an output force. Is this what is meant? If so, where is the fulcrum and what are the input and output forces?

ii. I don't understand what is being described in the second part of this sentence.

Line 148 "Intriguingly, there are four constructs exhibit unaffected or little enhanced ATPase hydrolysis but obstructed lipid transport"

Line 164 " Truncation of the CH does not change the completeness of the NBD"

What is meant by the completeness of the NBD?

line 194 – what precisely do these dimensions (42 Å or 40 Å) describe? – this implies that the TMDs have similar dimensions to accommodate their substrates – but lines 203 (and the abstract) state that the binding pocket of ALDP is much larger than ABCB4. Please clarify.

What are the volumes of the binding cavities in ALDP and ABCB4, and how does the volume of PC compare to the volume of the VLCFA substrate?

line 264 – Figure 4 – which PDBs are which? (ie ligand-free, substrate-bound, ATP-bound)?

One last point that likely reflects an issue on my end, but I was not able to see the EM map postrun9.mrc in coot 0.9.6 – Coot opened the map and provided a histogram of the density, but it didn't display anything over the "A_fit_real_space_refined-coot-34_real_space_refined_001_all_states.pdb" coordinates.

Reviewer #3 (Recommendations for the authors):

The authors have significantly improved the manuscript with the addition of additional functional data of several mutants.

A few more points to strengthen the manuscript:

Why is the structure in line 190 compared to ABCB4 and not the other structures listed in line 90? It makes more sense to discuss the mechanism based on the available structures of the same family in addition to the ABCB4.

It is still unclear how they selected their mutants; lines 122-125 do not have any refs of previously identified mutations. Do we know any of them being specific disease related?

Lines 85-91 contradict each other; they state that there is lipid-like density but at the same time they write that they have an apo structure.

The grammar in several sections needs to be improved.

https://doi.org/10.7554/eLife.75039.sa1

Author response

Essential revisions:

1. The modelled two putative Co-A esters lack the positively charged residues for coordinating phosphates. Indeed, the Co-A binding in the human ABCD1 structure by Chen and co-workers make more chemical sense https://www.biorxiv.org/content/10.1101/2021.09.24.461565v2.full.pdf. This study found they same shaped density as you have seen in your ABCD1 structure, but have modelled the lipid 18:0 Lyso PE molecule into this location instead. Judging the deposited maps, we think that this is the more likely lipid for your density, rather than CoA as proposed. We need to see an update of the substrate-binding site and the corresponding mutations to either the new substrate-binding site position as in the competing paper or more solid experimental data for keeping with the currently modelled substrate-binding site.

Point accepted. We have removed the description of the potential binding site as we originally proposed to avoid controversial interpretation on the binding site of VLCFA-CoA.

We compared all the reported structures of ABCD1. Wang et al. described the oleoyl-CoA bound structure. Chen and co-workers present the C22-CoA bound structure. Xiong et al. reported the C26 and C26-CoA bound conformation. Judged from the released structure of the oleoyl-CoA or C22:0-CoA bound ABCD1 and the figure of C26:0-CoA bound ABCD1 from Xiong et al. The lipid density in our model partially overlaps with the CoA group of oleoyl-CoA and C26:0-CoA and the long tail of C22:0-CoA. Among the many tail-like densities surrounding the transmembrane region of ALDP. The density we observed that inserted into the crevice of TMDs is relatively strong. This indicates an endogenous lipid may partially occupy the large binding cavity for the VLCFA-CoA and may provide explanations for the impaired lipid transport by the mutations around the cleft mentioned in our original manuscript.

Author response image 1

2. Certain assumptions have been made but there is no references or experimental validation from this work to warrant such conclusions; eg p6 'Pathogenic mutations on these residues may disturb the stability of ALDP. The residues of the third group are located on the NBDs, which may hinder the hydrolysis of nucleotides. Given the explosion of structural information on ABC transporters (for example, of the ~50 human ABC transporters, structures have been reported for the human A1, A4, B1, B2, B3, B4, B6, B8, B10, B11, C7, D4, G2, G5, G8 transporters in addition to D1), it is a challenge to interpret new results in the context of existing knowledge about ABC transporters. The transport assays provide an approach to connecting structure and function, but too little information is provided about the assays to make those connections. Of the ~900 mutants, why were the variants in figure 4B selected? Where are they located in the structure? What do the results of the transport assay mean? For example, is the reduction of C22:S-CoA in the cytosol for the G343V mutant due to impaired transport or to reduced expression, misfolding or some other cause? mutations are mainly mapped to the TMDs, which may affect the conformational changes during substrate transport.

Please show the correct trafficking, expression (normalised to WT) and folding of all mutations tested in HEK293 cells.

Please show that the purified protein is capable of turning over ATP and that the basal ATPase activity stimulated by the substrate. We would recommend carrying out complementary ATPase assays on key mutations also. It is critical to better integrate the mutagenesis data with the structure.

Point accepted. We have discussed the transport mechanism of ABC transporters in the revision based on integrative analysis of expression level, ATPase assay and trafficking. We also cited previously literatures that reported ALD-related mutants mentioned in our study. We mapped the mutations selected for transport assay on the structure as suggested. In the original manuscript, we temporarily classified the numerous ALD-associated mutations into four groups based on structural mapping of these mutations (Figure-3). Due to the complexity for preparing the lipid extraction to fulfill the transport assay, we further selected several representative mutations following the structural guidance. Additionally, we removed the peroxisomal helix and the C-terminal coil-coiled domain in ABCD1 to investigate the role of these two featured structural elements.

To interpretate the results of the transport assay, we analyzed the trafficking, expression, and the folding of these mutants through confocal microscope, quantitative Western blot, and thermo-stability assessment as the reviewers suggested. The ATPase activity of these mutations has been presented as well. (See Author response image 2)

Author response image 2

3. Please provide some rationale basis for the additional structural features found in ABCD1 (peroxisomal helix and the C-terminal coil-coiled domain) and how this might be related to its function?

As described above, we have individually truncated the peroxisomal helix and the C-terminal coil-coiled domain and examined the trafficking, expression, and the thermo-stability of these two constructs. The ATPase activity and the transport activity of these mutations has been presented as well. Judged from the ATPase and transport activity, the peroxisomal helix exhibit obstructed ATP hydrolysis and lipid transport (decreased ~50%), whereas deletion of the C-terminal coil-coiled domain have more influences on the lipid transport (decreased ~80%) than ATP hydrolysis. One possible reason is that deletion of the C-terminal domain has influence on the stability of ABCD1 with the NBD domain unaffected. Another possibility is that the C-terminal domain facilitates the conformational change of ABCD1 during the lipid transport. This conjecture is supported by structural studies reveals that the coil-coiled domain cannot be identified from the ATP bound state.

Author response table 1
StateConfirmationPeroxisomal helixc-terminal coiled domainReference
ApoCytosol-facingPresencePresenceThis study
Apo (chimera)Cytosol-facingAbsencePresenceChen et al
C22-CoA bound (chimera)IntermediatePresenceAbsenceChen et al
ATP bound (Chimera)Peroxisomal-facingAbsenceAbsenceChen et al
Oleoyl-CoA boundCytosol-facingPresencePresenceWang et al
ATP boundPeroxisomal-facingAbsenceAbsenceWang et al
C26 boundCytosol-facingPresencePresenceXiong et al
C26-CoA & ATP boundCytosol-facingPresencePresenceXiong et al
ATP boundPeroxisomal-facingAbsenceAbsenceXiong et al
ATP-γS bound state 1Peroxisomal-facingAbsenceAbsenceLe et al
ATP-γS bound state 2Cytosol-facingPresenceAbsenceLe et al

The peroxisomal helix becomes disordered in the presence of ATP. This conformational change led to the lipid-binding cavity open to the peroxisome lumen. The peroxisomal helix connects TM5 and TM6. TM5 is coordinated by the NBD of another protomer whereas TM6 is proceeded by NBD. Deletion of the peroxisomal helix may lead to spatial constraints for the movement of TM5 and TM6 towards the symmetry axis, finally transduced to the NBD.

Author response image 3

4. The final model is difficult to follow and this, of course, relates to the "actual" location of the substate-binding site. Is an outward-facing conformation to be expected in ABCD1? Furthermore, the conformational changes as proposed by comparing to AlphaFold models should be made with caution until experimental validation of these states. Please incorporate the AlphaFold models more carefully into the paper, i.e.,what are the state-specific co-evolved residue-pairs predicted in the AlphaFold models, do these make sense, and can these contact pairs be experimentally validated by mutagenesis?

As we mentioned above (major point 1), we have rewritten the description of the potential binding site in our model to avoid controversial interpretation on the binding site of VLCFA-CoA. We found that the lipid density in our model partially overlaps with the acyl tail of C22-CoA, identified by Chen and co-workers. Further, we compared our structure and the reported ALDP models. Our structure is similar to the apo-state by Chen et al. and the Oleoyl-CoA bound state by Wang et al. Similarly, both the apo and oleoyl-CoA bound state exhibit cytosolic-facing conformation. Based on the structure alignment and functional state of our structure, we consider our state is cytosolic-facing conformation.

For the AlphaFold models in the original paper, we used the model from the AlphaFold Protein Structure Database for structure comparison. The database provides monomer structure model rather than multimer. As suggested by the reviewers, we provide the dimeric model through AlphaFold predicted on local-computer (see Author response image 4). This process follows the general command of AlphaFold by providing the sequence file. However, we are inaccessible to read out the state-specific co-evolved residue-pairs used by AlphaFold at the current stage. Instead, we performed the co-evolved residues by another web-server developed by David Baker group (http://gremlin.bakerlab.org/). We provided the results along with the revised manuscript. More co-evolved residues pairs are located at the NBDs than TMDs (see Author response image 5).

Author response image 4
Author response image 5

The predicted ALDP-dimer is reminiscent of the ATP-bound state. Such result is not surprising because the majority conformation of multiple ABC transporters has been determined to be the ATP-bound states [ref]. Before the cryo-EM structures of ALDP from the several independent groups, the closest homolog of ABCD1 with structure is ABCD4, which displays an ATP-bound conformation. Structural alignment between our model and the predicted or the ATP-bound state reveals dramatic conformational changes at the TMD and NBD.

5. The paper needs to be extensively re-written. We would like to see how the structure uncovers new mechanistic insights. With the additional functional data requested, consider how the ABCD1 structure enables a understanding of why very long chain fatty acids use such an ABC transporter structures for being imported into peroxisomes. How does this differ to other ABC transporters that transport lipids and why?

We compared the differences of the binding cavities between ALDP and ABCB4. ABCB4 is a phosphotidylcholine transporter. Typically, ALDP consists of two binding pockets listed in the large cavity formed by TMDs. ABCB4 harbors one lipid binding pocket. The cavity of ALDP is larger than ABCB4. The different binding manner of these two ABC transporters is associated with distinct conformational changes during lipid transport. The two protomers of ALDP moves away from the cavity center in the presence of ATP. These differences may be attributed to two reasons: (1) ALDP is assembled by two identical protomers whereas ABCB4 is assembled by single chain polypeptides; (2) The chemical shape of the specific substrates. Unlike most fatty acids, VLCFAs are too long to be metabolized in the mitochondria and must be metabolized in peroxisomes. We use C22:0 lipids for the comparison. C22:0-CoA has one acyl tail and a much larger hydrophilic head in contrast to C22:0-PC.

Author response image 6

Reviewer #1 (Recommendations for the authors):

1. The modelled two putative Co-A esters lack the positively charged residues for coordinating phosphates. Indeed, the Co-A binding in the human ABCD1 structure by Chen and co-workers make more chemical sense https://www.biorxiv.org/content/10.1101/2021.09.24.461565v2.full.pdf. This study found they same shaped density as you seen in your ABCD1 structure, but have modelled the lipid 18:0 Lyso PE molecule into this location instead. Judging the deposited maps I think that this is the more likely lipid for your density, rather than CoA as proposed. Please read this submission more carefully and update accordingly as I find it worrying that this has not been picked up.

Point accepted. Please see “Essential Revisions (for the authors)” Major point 1.

2. Please check that ABCD1 mutations are trafficked correctly to peroxisomes. Also one needs to normalise uptake against expression level. Please complement substrate uptake experiments with ATPase activity measurements also.

Point accepted. Please see “Essential Revisions (for the authors)” Major point 2.

3. This paper is difficult to follow in places and using AlphaFold structures to make conclusions regarding local conformational changes seems premature at this stage without further validation.

Please see “Essential Revisions (for the authors)” Major point 4.

4. Please provide some rationale basis for the additional structural features found in ABCD1 (peroxisomal helix and the C-terminal coil-coiled domain) and how this might be related to its function?

Please see “Essential Revisions (for the authors)” Major point 3.

We thank this reviewer for his/her thoughtful suggestions which we fully accepted.

Reviewer #3 (Recommendations for the authors):

The CoA-ester binding should be discussed in more details.

Although several mutants have been prepared, they are not really discussed in the context of the manuscript or previously published work. The discussion needs to expand further. Certain assumptions have been made but there is no references or experimental validation from this work to warrant such conclusions; eg p6 'Pathogenic mutations on these residues may disturb the stability of ALDP. The residues of the third group are located on the NBDs, which may hinder the hydrolysis of nucleotides. The other mutations are mainly mapped to the TMDs, which may affect the conformational changes during substrate transport.'. Based on their structure it should be possible to explain better why and how the mutants may affect the function.

The authors should show that the purified protein is capable of turning over ATP and that the basal ATPase activity stimulated by the substrate.

Point accepted. We have discussed the mutants in the revision based on the supplement of ATPase activity assay, trafficking information, thermo-stability assessment.

Figure 4B; are all mutants expressing at similar levels as the WT protein? They authors should show a Western Blot. Additionally, what do the error bars represent?

Point accepted. We have shown the WB results of the expression level in the revision.

The authors should briefly state in the M and M how they prepared the Alphafold models; ie collar or full Alphafold. Why not use the Alphafold multimer and make the homodimer?

Please see “Essential Revisions (for the authors)” Major point 4.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

– The main impact of the paper is the ALDP structure with a number of supporting mutations. The current Manuscript word length is around 1800 words and 4 figures, which is close to the requirements for a short report (1500 words and 3 figures). We think the paper would come across much better as a short report, and ask that you please re-write to meet these requirements and improve the readability of the text in the process.

Thanks for the kind suggestions. To comply with your editorial requests, the revised manuscript is within 1500 words and 3 figures. We thoroughly rewritten the main text following the review comments and suggestions.

The referees have further concerns to be addressed:

– line 100: "The TM2 serves as a mechanical lever: the cytosolic side moves towards the central cavity, while the peroxisomal side may move backwards the symmetry axis"

Two points:

i. is TM2 really a mechanical level? A lever is a rigid object with a fulcrum point that amplifies an input force to become an output force. Is this what is meant? If so, where is the fulcrum and what are the input and output forces?

ii. I don't understand what is being described in the second part of this sentence.

Point accepted. We have removed the “mechanical level” as this description is not accurate.

– Line 148 "Intriguingly, there are four constructs exhibit unaffected or little enhanced ATPase hydrolysis but obstructed lipid transport"

To discriminate the disease-derived mutants and the rational designed mutants, we have rewritten this paragraph and removed the subjective vocabulary like “Intriguingly”. Please see page 7 in the revised manuscript.

– Line 164 " Truncation of the CH does not change the completeness of the NBD"

What is meant by the completeness of the NBD?

We keenly accept the suggestion from this reviewer and rewrite this sentence as below:

“Truncation of CH (Δ683-745) does not change the assembly of NBD but affects the conformational changes between the two protomers”

– Line 194 – what precisely do these dimensions (42 Å or 40 Å) describe? – this implies that the TMDs have similar dimensions to accommodate their substrates – but lines 203 (and the abstract) state that the binding pocket of ALDP is much larger than ABCB4. Please clarify.

What are the volumes of the binding cavities in ALDP and ABCB4, and how does the volume of PC compare to the volume of the VLCFA substrate?

Point accepted. The dimensions of 42 Å or 40 Å imply the TMDs have similar dimensions. But ALDP has two substrate binding pockets and two exits for substrate release, while ABCB4 coordinates one PC molecule. In the revised manuscript, we have rewritten this part. Because the both VLCFA and PC are flexible, we temporarily compared the molecule size through describing the molecular weight and roughly measured the length of these two molecules in revised Figure 3 —figure supplement 5. The volumes of one binding cavity is ~1695 Å3 for C22:0-CoA and ~1236 Å3 for PC.

– Line 264 – Figure 4 – which PDBs are which? (ie ligand-free, substrate-bound, ATP-bound)?

Point accepted. We have labeled the states of corresponding PDB files.

– One last point that likely reflects an issue on my end, but I was not able to see the EM map postrun9.mrc in coot 0.9.6 – Coot opened the map and provided a histogram of the density, but it didn't display anything over the "A_fit_real_space_refined-coot-34_real_space_refined_001_all_states.pdb" coordinates.

We have provided the map and coordinate again in this revision. Please see attached files.

– Why is the structure in line 190 compared to ABCB4 and not the other structures listed in line 90? It makes more sense to discuss the mechanism based on the available structures of the same family in addition to the ABCB4.

Point accepted. We have compared the structures of ALDP in different conformations as suggested. In the last version, we wanted to compare the ABC transporters of different long-tail lipids, thus we selected the structures of ALDP and ABCB4. This part has been moved to the supplementary figures.

– It is still unclear how they selected their mutants; lines 122-125 do not have any refs of previously identified mutations. Do we know any of them being specific disease related?

Point accepted. We have rewritten this part to describe how these mutants have been selected. The references have been cited properly as suggested. Briefly, we selected six ALD-derived mutants that localized at the substrate binding site, the NBD and the C-terminal coiled-coil domain. These mutants are W339R, S342P, G343V, A396T, Q544R, and T693M. Additionally, we designed four constructs based on the structural findings, including S164A/Y310A, E291A/R518A, Δ364-374 to delete the peroxisomal helix and Δ683-745 to truncate the whole C-terminal coiled-coil domain. These four mutants are designed for probing how the associated structure elements affect the activity of ALDP.

– Lines 85-91 contradict each other; they state that there is lipid-like density but at the same time they write that they have an apo structure.

Point accepted. We compared the structures of ALDP from different groups. Chen et al. also reported two lipid-like densities in the apo-state which have been built as PE in their model. Nonetheless, to avoid inaccurate interpretations, we have change the description as unknown densities.

Reviewer #2 (Recommendations for the authors):

The structure of human ALDP is of significance given the role of this transporter in the transport of VLCFAs across the peroxisomal membrane. Nearly a thousand ALD-causing mutations have been identified for ALDP and the structure provides a framework for assessing their functional consequences. In this revised manuscript, the authors have addressed many of the reviewers' concerns in the rebuttal letter. Given the importance of this transporter, particularly the human version of ALDP, I am sympathetic to the publication.

The manuscript, however, needs extensive polishing to be suitable for publication, however. Here are a few examples

line 100: "The TM2 serves as a mechanical lever: the cytosolic side moves towards the central cavity, while the peroxisomal side may move backwards the symmetry axis"

Two points:

i. is TM2 really a mechanical level? A lever is a rigid object with a fulcrum point that amplifies an input force to become an output force. Is this what is meant? If so, where is the fulcrum and what are the input and output forces?

ii. I don't understand what is being described in the second part of this sentence.

Please see above.

Line 148 "Intriguingly, there are four constructs exhibit unaffected or little enhanced ATPase hydrolysis but obstructed lipid transport"

Please see above.

Line 164 " Truncation of the CH does not change the completeness of the NBD"

What is meant by the completeness of the NBD?

Please see above.

line 194 – what precisely do these dimensions (42 Å or 40 Å) describe? – this implies that the TMDs have similar dimensions to accommodate their substrates – but lines 203 (and the abstract) state that the binding pocket of ALDP is much larger than ABCB4. Please clarify.

What are the volumes of the binding cavities in ALDP and ABCB4, and how does the volume of PC compare to the volume of the VLCFA substrate?

Please see above.

line 264 – Figure 4 – which PDBs are which? (ie ligand-free, substrate-bound, ATP-bound)?

Please see above.

One last point that likely reflects an issue on my end, but I was not able to see the EM map postrun9.mrc in coot 0.9.6 – Coot opened the map and provided a histogram of the density, but it didn't display anything over the "A_fit_real_space_refined-coot-34_real_space_refined_001_all_states.pdb" coordinates.

Please see above.

We thanks this reviewer for his/her constructive suggestions.

Reviewer #3 (Recommendations for the authors):

The authors have significantly improved the manuscript with the addition of additional functional data of several mutants.

A few more points to strengthen the manuscript:

Why is the structure in line 190 compared to ABCB4 and not the other structures listed in line 90? It makes more sense to discuss the mechanism based on the available structures of the same family in addition to the ABCB4.

Please see above.

It is still unclear how they selected their mutants; lines 122-125 do not have any refs of previously identified mutations. Do we know any of them being specific disease related?

Please see above.

Lines 85-91 contradict each other; they state that there is lipid-like density but at the same time they write that they have an apo structure.

Please see above.

The grammar in several sections needs to be improved.

We have rewritten the manuscript and examined the grammars. We keenly thank this reviewer for his/her constructive suggestions.

https://doi.org/10.7554/eLife.75039.sa2

Article and author information

Author details

  1. Yutian Jia

    State Key Laboratory for Agrobiotechnology, Department of Nutrition and Health, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Data curation, Formal analysis, Investigation
    Contributed equally with
    Yanming Zhang
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6921-7222
  2. Yanming Zhang

    State Key Laboratory for Agrobiotechnology, Department of Nutrition and Health, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Data curation, Formal analysis, Investigation
    Contributed equally with
    Yutian Jia
    Competing interests
    No competing interests declared
  3. Wenhao Wang

    State Key Laboratory for Agrobiotechnology, Department of Nutrition and Health, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Data curation, Formal analysis, Investigation
    Competing interests
    No competing interests declared
  4. Jianlin Lei

    Technology Center for Protein Sciences, Ministry of Education Key Laboratory of Protein Sciences, School of Life Sciences, Tsinghua University, Beijing, China
    Contribution
    Software
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-9384-8742
  5. Zhengxin Ying

    State Key Laboratory for Agrobiotechnology, Department of Nutrition and Health, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Resources, Formal analysis, Investigation
    For correspondence
    yingzhengxin@cau.edu.cn
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0948-4948
  6. Guanghui Yang

    State Key Laboratory for Agrobiotechnology, Department of Nutrition and Health, College of Biological Sciences, China Agricultural University, Beijing, China
    Contribution
    Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    guanghuiyang@cau.edu.cn
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-6835-1611

Funding

National Natural Science Foundation of China (3217110084)

  • Guanghui Yang

National Natural Science Foundation of China (32130081)

  • Guanghui Yang

National Natural Science Foundation of China (32100759)

  • Zhengxin Ying

Chinese Universities Scientific Fund (15050004)

  • Guanghui Yang

Chinese Universities Scientific Fund (15051002)

  • Guanghui Yang

Chinese Universities Scientific Fund (15050017)

  • Guanghui Yang

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank the Tsinghua University Branch of China National Center for Protein Sciences (Beijing) for the cryo-EM facility and the computational facility support, Dr. Xiaomin Li, Dr. Fan Yang and Tao Liu for technical support in EM data acquisition. We thank Mr. Wei Gao and Ms. Yang Yue at Mo-In Biotechnology (Beijing) Co., Ltd for the data acquisition and analysis of the protein thermostability. This work was supported by National Natural Science Foundation of China (3217110084; 32130081; 32100759); Chinese Universities Scientific Fund (15050004; 15050017; 15051002); and Young Elite Scientists Sponsorship Program by China Association for Science and Technology.

Senior Editor

  1. Kenton J Swartz, National Institute of Neurological Disorders and Stroke, National Institutes of Health, United States

Reviewing Editor

  1. David Drew, Stockholm University, Sweden

Reviewers

  1. David Drew, Stockholm University, Sweden
  2. Konstantinos Beis

Publication history

  1. Preprint posted: September 25, 2021 (view preprint)
  2. Received: October 28, 2021
  3. Accepted: November 10, 2022
  4. Accepted Manuscript published: November 14, 2022 (version 1)
  5. Version of Record published: November 23, 2022 (version 2)

Copyright

© 2022, Jia, Zhang et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 451
    Page views
  • 139
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Yutian Jia
  2. Yanming Zhang
  3. Wenhao Wang
  4. Jianlin Lei
  5. Zhengxin Ying
  6. Guanghui Yang
(2022)
Structural and functional insights of the human peroxisomal ABC transporter ALDP
eLife 11:e75039.
https://doi.org/10.7554/eLife.75039

Further reading

    1. Structural Biology and Molecular Biophysics
    Jasenko Zivanov, Joaquín Otón ... Sjors HW Scheres
    Tools and Resources

    We present a new approach for macromolecular structure determination from multiple particles in electron cryo-tomography (cryo-ET) data sets. Whereas existing subtomogram averaging approaches are based on 3D data models, we propose to optimise a regularised likelihood target that approximates a function of the 2D experimental images. In addition, analogous to Bayesian polishing and contrast transfer function (CTF) refinement in single-particle analysis, we describe approaches that exploit the increased signal-to-noise ratio in the averaged structure to optimise tilt series alignments, beam-induced motions of the particles throughout the tilt series acquisition, defoci of the individual particles, as well as higher-order optical aberrations of the microscope. Implementation of our approaches in the open-source software package RELION aims to facilitate their general use, in particular for those researchers who are already familiar with its single-particle analysis tools. We illustrate for three applications that our approaches allow structure determination from cryo-ET data to resolutions sufficient for de novo atomic modelling.

    1. Structural Biology and Molecular Biophysics
    Metehan Ilter, Ramazan Kaşmer ... Ozge Sensoy
    Research Article

    Undruggability of RAS proteins has necessitated alternative strategies for the development of effective inhibitors. In this respect, phosphorylation has recently come into prominence as this reversible post-translational modification attenuates sensitivity of RAS towards RAF. As such, in this study, we set out to unveil the impact of phosphorylation on dynamics of HRASWT and aim to invoke similar behavior in HRASG12D mutant by means of small therapeutic molecules. To this end, we performed molecular dynamics (MD) simulations using phosphorylated HRAS and showed that phosphorylation of Y32 distorted Switch I, hence the RAS/RAF interface. Consequently, we targeted Switch I in HRASG12D by means of approved therapeutic molecules and showed that the ligands enabled detachment of Switch I from the nucleotide-binding pocket. Moreover, we demonstrated that displacement of Switch I from the nucleotide-binding pocket was energetically more favorable in the presence of the ligand. Importantly, we verified computational findings in vitro where HRASG12D/RAF interaction was prevented by the ligand in HEK293T cells that expressed HRASG12D mutant protein. Therefore, these findings suggest that targeting Switch I, hence making Y32 accessible might open up new avenues in future drug discovery strategies that target mutant RAS proteins.