Substrate Specificity and the Direction of Transport in the ABC Transporters ABCD1–3 and ABCD4

Abstract

The ATP-binding cassette (ABC) transporters are one of the largest families of membrane-bound proteins and exist in almost all living organisms from eubacteria to mammals. They transport diverse substrates across membranes utilizing the energy of ATP hydrolysis as a driving force and play an essential role in cellular homeostasis. In humans, four ABC transporters classified as subfamily D have been identified. ABCD1–3 are localized to peroxisomal membranes and involved in the transport of various acyl-CoAs from the cytosol to the peroxisomal lumen. ABCD4 functions on the lysosomal membranes and transports vitamin B₁₂ (cobalamin) from lysosomes into the cytosol. The mutation of genes encoding ABCD1, ABCD3, and ABCD4 are responsible for genetic diseases called X-linked adrenoleukodystrophy, congenital bile acid synthesis defect 5, and cobalamin deficiency, respectively. In this review, we summarize the targeting mechanism and physiological functions of the ABCD transporters and discuss insights that have been obtained on the transport mechanism based on disease-causing mutations and cryo-electron microscopy (EM) structural studies.

1. Introduction

ATP-binding cassette (ABC) transporters comprise a superfamily of membrane-bound proteins found in almost all organisms from eubacteria to mammals. ABC transporters share a common domain architecture consisting of two nucleotide-binding domains (NDBs) that hydrolyze ATP and two transmembrane domains (TMDs) that form a translocation pathway.¹⁾ They catalyze the ATP-dependent transmembrane transport of a wide variety of substrates and are categorized into two types: exporters that transport substrate from the cytosol to the extracellular space or the lumen of subcellular organelles, and importers that transport substrates into the cytosol.²⁾ Prokaryotic ABC transporters include both exporters and importers.^3,4) On the other hand, almost all eukaryotic ABC transporters are exporters, with only a few exceptions in plants and mammals.^5–9)

The human ABC transporter family at present comprises 48 members that are classified into seven subfamilies, A to G, based on architecture and amino acid homology.¹⁰⁾ Members of ABCE and ABCF have no TMDs and are involved in processes unrelated to transport.^11,12) The remaining families are categorized into two groups according to their TMD architecture and auxiliary domains. One contains ABCB, ABCC and ABCD, and possesses long transmembrane helices that extend into the cytosol. The other contains ABCA and ABCG, which are composed of shorter transmembrane helices, thus bringing the NBDs closer to the membrane.^2,13) ABC transporters exist on plasma membranes and subcellular compartments such as the endoplasmic reticulum (ER), peroxisomes, mitochondria, lysosomes, and Golgi apparatus, and play essential roles in maintaining cellular homeostasis. Their dysfunction is related to various diseases.¹⁴⁾

In humans, four ABCD transporters have been identified: ABCD1 (adrenoleukodystrophy protein/ALDP), ABCD2 (ALDP-related protein/ALDRP), ABCD3 (the 70-kDa peroxisomal membrane protein/PMP70) and ABCD4 (the PMP70- related protein/P70R).^15–18) The ABCD transporters are an ABC half-transporter, with one TMD and one NBD, and mainly function as a homodimer¹⁹⁾ (Fig. 1), but a heterodimeric structure has also been suggested for ABCD1–3.²⁰⁾ ABCD1–3 are peroxisomal proteins and play an important role in the transport of various acyl-CoAs into peroxisomes for further β-oxidation. ABCD1 and ABCD2 are suggested to be involved in the transport of long chain fatty acid (LCFA)-CoA and very long chain fatty acid (VLCFA)-CoA with different substrate specificities, while ABCD3 is involved in the transport of dicarboxylic acid, branched chain acyl-CoA and the bile acid intermediates, di- and tri-hydroxycholestanoyl (DHCA and THCA)-CoA.²¹⁾ The dysfunction of ABCD1 and ABCD3 is the respective cause of X-linked adrenoleukodystrophy (X-ALD), a neurodegenerative disease, and congenital bile acid synthesis defect (CBAS) type 5, a liver disease with hepatosplenomegaly.^16,22) On the other hand, ABCD4 is located on the lysosomal membrane.²³⁾ Since the dysfunction of ABCD4 is the cause of vitamin B₁₂ (cobalamin) deficiency with the phenotype of a failure to release cobalamin from lysosomes, ABCD4 is considered to be involved in the transport of cobalamin from the lysosomal lumen to the cytosol.²⁴⁾ Recently, we demonstrated that ABCD4 reconstituted into liposomes is capable of transporting cobalamin from the inside to the outside of liposomes in an ATPase activity-dependent manner.²⁵⁾ In addition, it is predicted that ABCD4 possesses the topological feature where the COOH-terminal half, including the NBD, is exposed to the cytosol.²⁶⁾ These strongly suggest that ABCD4 is able to transport cobalamin from the inside of lysosomes to the cytosol; that is, ABCD4 functions as an importer.

Fig. 1. Schematic Structure of the ABCD Transporters

ABCD transporters are half-size ABC protein with one transmembrane domain (TMD) and one nucleotide-binding domain (NBD), and mainly function as a homodimer. Two TM helices (TM4 and 5) from the TMD are swapped into the other TMDs. Abbreviations: N, N terminus; C, C terminus; CH, coupling helix; NBD, nucleotide-binding domain; TMD, transmembrane domain.

Although ABCD transporters share a high homology of amino acid sequence and have the same membrane topology, ABCD1–3 and ABCD4 have distinct features. In this review, we focus on the differences in subcellular localization, substrate specificity and transport direction.

2. The Targeting Mechanism of ABCD Proteins

The subcellular localization of ABCD1–3 and ABCD4 are peroxisomes and lysosomes, respectively. The ABCD transporters are translated on cytosolic free polysomes, and then their trafficking branches, depending on whether they possess the NH₂-terminal hydrophobic segment (H0 motif) which is adjacent to the first transmembrane helix (TM1) or not²⁷⁾ (Fig. 2).

Fig. 2. Subcellular Localization of the ABCD Transporters

The NH₂-terminal H0 motif plays a key role in the trafficking. ABCD1–3, which possess the H0 motif, are recognized by Pex19p and then integrated into the peroxisomal membrane. In contrast, ABCD4 lacks the H0 motif and is captured by signal recognition particles (SRP) instead of Pex19p, and destined to the ER. Subsequently, ABCD4 is translocated to lysosomes in a manner dependent on the lysosomal membrane protein LMBD1.

ABCD1–3 containing the H0 motif are selectively captured by membrane biogenesis factor Pex19p, which plays a critical role in the targeting of peroxisomal membrane proteins, and then are inserted into the peroxisomal membrane. Among the peroxisomal ABCD transporters, the trafficking of newly synthesized ABCD3 has been studied in the most detail.^28–31) As ABCD3 is an integral membrane protein with high hydrophobicity, forming a complex with Pex19p is indispensable for ABCD3 to maintain a proper conformation for targeting peroxisomes. Pex19p is a chaperon-like protein and inserts ABCD3 into the peroxisomal membrane via an interaction with docking receptor Pex3p on the membrane. To investigate the critically important region of ABCD3 for targeting peroxisomes, we prepared various mutated ABCD3 fused with green fluorescent protein (GFP) and examined their subcellular localization in Chinese hamster ovary (CHO) cells. The analysis clarified that ABCD3 is recognized and interacts with Pex19p via the H0 motif and the region of TM5–TM6. In addition, the pairs with hydrophobic amino acid isoleucine (Ile)⁷⁰-leucine (Leu)⁷¹ and Ile³⁰⁷-Leu³⁰⁸ adjacent to TM1 and TM5, respectively, might be peroxisomal membrane protein targeting signals (mPTSs). Thus, the complex of ABCD3 and Pex19p is transported to peroxisomes depending on the mPTSs of ABCD3. Finally, ABCD3 is integrated into the peroxisomal membrane. Furthermore, it was demonstrated that TM1 of ABCD3 possesses strong ER targeting ability and the NH₂-terminal short motif composed of nine amino acid residues functions as a cis-acting element suppressing ER targeting. Among these amino acid residues, serine (Ser)⁵ is indispensable for suppression.³²⁾

In the case of ABCD1, it was shown that the amino acid region at positions 68–82, which is adjacent to the deduced TM1, is Pex19p-binding site and coincides with its mPTS.³³⁾ Particularly, the three amino acids Leu⁷⁸-Leu⁷⁹-arginine (Arg)⁸⁰ have a critical role in the targeting of ABCD1 to peroxisomes. These amino acids are equivalent of Ile⁷⁰-Leu⁷¹-lysine (Lys)⁷² in ABCD3, which are revealed as an mPTS in ABCD3. In ABCD2, the amino acid region at positions 84–97 is deduced to be a Pex19p-binding site.³³⁾ However, there is no experimental data yet.

On the other hand, ABCD4 is captured by certain signal recognition particles instead of Pex19p and delivered to the ER since ABCD4 lacks the H0 motif.²⁶⁾ Subsequently, the subcellular localization of ABCD4 is shifted to lysosomes through forming a complex with the lysosomal membrane protein LMBD1.²³⁾ We clarified the underlying mechanism of this translocation. ABCD4 is localized in ER membranes as a homodimer when it is expressed alone in HuH7 cells. The localization of ABCD4 is drastically changed to lysosomes when LMBD1 is co-expressed in the cells. In the course of this translocation, the ABCD4 dimer interacts with LMBD1, as demonstrated by pull-down assay. When co-expressed with mutated LMBD1 at a putative AP-2 binding motif that traffics to the plasma membrane, ABCD4 is also mislocalized to the plasma membrane. These results indicate that the lysosomal targeting ability of LMBD1 is indispensable for the translocation of ABCD4 from the ER to lysosomes. In addition, it was demonstrated that endogenous ABCD4 is localized to both lysosomes and the ER in HEK293 cells by density gradient centrifugation with iodixanol, and it was shown to form a complex with LMBD1 by co-immunoprecipitation. Furthermore, ABCD4 in the lysosomal fraction was reduced to approx. 40% and ABCD4 in the ER fraction was increased in LMBD1 knockout HEK293 cells compared with wild type cells, suggesting that LMBD1 also has responsibility for the lysosomal localization of endogenous ABCD4.

3. Function of ABCD Proteins

Peroxisomes are organelles found in almost all eukaryotic cells and indispensable for cellular homeostasis. Their indispensable functions include the β-oxidation of fatty acids, especially VLCFA (>C22), as well as the synthesis of bile acid and plasmalogen. ABCD1–3 are involved in the transport of metabolites into the peroxisomal lumen (Fig. 3). ABCD1 dysfunction is responsible for X-ALD, which has a characteristic feature of the abnormal accumulation of VLCFA in tissues. It has been demonstrated that, in X-ALD skin fibroblasts, reduced VLCFA β-oxidation is recovered by exogenous expression of ABCD1, and the elevated VLCFA content is sequentially reduced to normal.³⁴⁾ In fact, ABCD1 has been shown to have substrate specificity for saturated, monounsaturated, and polyunsaturated VLCFA-CoA, such as C18:0-, C22:0-, C24:0-, C26:0-, C18:1-, and C24:6-CoA, using human ABCD1 expressing pxa1/pxa2Δ yeast cells, which lack endogenous peroxisomal ABC transporters.^35,36) Recently, it has been revealed that C26:1-CoA, not C26:0-CoA, is the most abundant among the VLCFA-CoA species in fibroblasts from X-ALD patients and in ABCD1-deficient HeLa cells.³⁷⁾

Fig. 3. Substrate Specificity of the ABCD Transporters

The substrate specificities of ABCD1 and ABCD2 are overlapped. However, ABCD1 has a higher specificity for VLCFA-CoAs, such as C24:0-CoA and C26:0-CoA, than ABCD2. In contrast, ABCD2 has an affinity for polyunsaturated fatty acyl-CoA, but ABCD1 does not. ABCD3 is suggested to be involved in the transport of LCFA-CoA, branched chain acyl-CoA and the bile acid intermediates such as THCA-CoA and DHCA-CoA. On the other hand, ABCD4 is involved in the cobalamin transport from lysosomal lumen to the cytosol.

ABCD2 is functionally redundant with ABCD1.³⁸⁾ In fact, exogenous expression of ABCD2 adequately restores the defect of β-oxidation in X-ALD fibroblasts. Using ABCD2 expressing pxa1/pxa2Δ yeast cells, ABCD2 displays substrate specificities that overlap with ABCD1 toward saturated and monounsaturated fatty acyl-CoAs, but has a lower specificity for saturated VLCFA-CoA. In contrast, ABCD2 has an affinity for polyunsaturated fatty acids such as C22:6-CoA and C24:6-CoA, but ABCD1 does not.³⁶⁾ However, no disease is reported due to ABCD2 dysfunction, since the level of endogenous ABCD2 is quite low in human cells.

ABCD3 is one of the most abundant proteins on peroxisomal membranes, at least in hepatocytes.³⁹⁾ ABCD3 shows overlap substrate specificity with ABCD1 and ABCD2, but also distinguishing specificity for the transport of dicarboxylic acids, branched-chain fatty acids, and C27 bile acid intermediates, DHCA and THCA-CoA, when overexpressed in pxa1/pxa2Δ yeast cells.⁴⁰⁾ The patients with an ABCD3 defect indeed exhibit a marked accumulation of C27 bile acid intermediates in plasma and the disease in humans is known as CBAS type 5.²²⁾

Lysosomes are organelles present in all eukaryotic cells. The organelles keep their internal pH around 5.0 and play a crucial role in a number of physiological processes.⁴¹⁾ ABCD4 is involved in the transport of cobalamin from the inside of lysosomes to the cytosol, since ABCD4 dysfunction results in failure of lysosomal cobalamin efflux.²⁴⁾ In the blood, cobalamin exists as a complex with transcobalamin, which is taken up into lysosomes through receptor-mediated endocytosis. Subsequently, cobalamin is released from the lysosomes as a free form and converted into methylcobalamin and adenosylcobalamin, which act as cofactors for methionine synthase and methylmalonyl-CoA mutase, respectively.⁴²⁾ As mentioned above, ABCD4 dysfunction results in the accumulation of cobalamin in the lysosomal lumen. Since a similar phenotype is caused by LMBD1 dysfunction,⁴³⁾ these two proteins are considered to function as a complex in the cobalamin transporting process from the lysosomal lumen to the cytosol. However, it has been shown that LMBD1 is necessary to translocate ABCD4 from the ER to lysosomes, and is not required for the cobalamin transport²³⁾ (See below).

There are several studies exhibiting the physical interaction between each of the ABCD transporters and other proteins.^44–47) Recently, several studies have shown that ABCD1 binds to proteins involved in lipid metabolism, protein/amino acid metabolism, peroxisomal matrix protein import, peroxisome organization, and membrane assembly.⁴⁷⁾ These results indicate the involvement of ABCD1 in other processes in addition to the transport of VLCFA-CoA into peroxisomes. Furthermore, ABCD4 has been shown to interact with the proteins, cyanocobalamin reductase/alkylcobalamin dealkylase and cobalamin trafficking protein CblD, that help cobalamin processing,⁴⁵⁾ suggesting that cobalamin is transported efficiently from lysosomes.

4. The Transport Mechanism of ABCD1 and ABCD4

Concerning the transport mechanism, ABCD1 is the most extensively studied of these transporters. However, the precise mechanism for transporting fatty acid through ABCD1 is controversial and there are two models. The first is that esterified fatty acids are directly transported into the peroxisomal matrix. The second is that free fatty acids (FAs) generated by the hydrolysis of acyl-CoAs are transported and then re-esterified by a peroxisomal acyl-CoA synthetase in the matrix. The first model is supported by the fact that the addition of CoA has no effect on the β-oxidation of VLCFA-CoA in isolated peroxisomes from human fibroblasts.⁴⁸⁾ On the other hand, studies in yeast models have shown that the CoA moiety is cleaved while transporting since ¹⁸OH from [¹⁸O] H₂O is incorporated into acyl-CoAs during incubation in yeast cells.⁴⁹⁾ In addition, COMATOSE, a homolog of human ABCD1 in Arabidopsis thaliana, has been shown to possess intrinsic acyl-CoA thioesterase (ACOT) activity, supporting the notion that hydrolysis of VLCFA-CoA occurs prior to transport.⁵⁰⁾ We have developed a procedure for expressing human ABCD1 in methylotrophic yeast, purifying ABCD1 and reconstituting the detergent-solubilized protein into liposomes,⁵¹⁾ and we analyzed ACOT activity and acyl-CoA transport. We determined the properties of ABCD1 ACOT activity using ABCD1 containing liposomes and 7-nitro-2-1,3-benzoxadiazol-4-yl (NBD)-labeled C16-CoA as a substrate.⁵²⁾ ABCD1 hydrolyzes NBD-C16-CoA and forms an intermemdiate with NBD-C16-CoA (acylation) during hydrolysis. Since ACOT activity and the acylation of ABCD1 are inhibited by p-chloromercuribenzoic acid, a cysteine-reactive compound, certain cysteine residue(s) are presumed to be involved in ACOT activity. In addition, it was shown that a truncated ABCD1 mutant containing only the NH₂-terminal TMD and missense ABCD1 mutant deficient in ATPase activity show ACOT activity comparable to the wild type ABCD1. ACOT activity is thus presumed to be located in the TMD and independent of ATPase activity. Subsequently, we established an assay for the transport of substrate into ABCD1-liposomes using NBD-C16-CoA, and directly demonstrated that ABCD1 transports free VLCFA into peroxisomes after the hydrolysis of VLCFA-CoA. During this transport process, the ACOT activity of ABCD1 is indispensable. This transport is thought to be composed of three steps; (i) the capture of VLCFA-CoA in the ACOT domain of ABCD1, (ii) acylation (covalent binding) of the VLCFA moiety to ABCD1, and (iii) the hydrolysis and release of VLCFA from ABCD1. The free CoA that is generated is transported via the ABCD transporters and released into the peroxisomal lumen, as demonstrated using peroxisomes isolated from Saccharomyces cerevisiae.⁵³⁾

Cryo-electron microscopy (EM) and X-ray crystallography have been provided insight into the substrate recognition and transport mechanisms. Most recently, the cryo-EM structures of the human ABCD1 transporters were reported including preprints in bioRχiv.^54–58) ABCD1 takes an inward-open configuration in the cytosol without any substrate or ATP (Fig. 4A). In the substrate-bound structure using C22:0-CoA as the substrate, the distance of each NBD become closer than the distance without substrate, and substrate-like densities are observed in the cavity between the TM helices, although the substrate is located in the central cavity between the TMDs in other ABC transporters.^59,60) Interestingly, two acyl-CoAs are suggested to be inserted deeply into the cavity. The acyl chain is located in a hydrophobic cleft as a cluster of hydrophobic residues crossing TM3–6. The CoA moiety is located in a pocket that is mainly composed of positively charged residues. The adenine ring is stabilized by hydrogen bonds with Ser²¹³ and Lys²¹⁷ on TM3, Arg⁴⁰¹ on TM6, and by π–π stacking with tryptophan (Trp)³³⁹ on TM6. The 3′- phosphate of ribose interacts with Arg¹⁵² on TM2, Gln^332′and Ser^340′ on TM5′, and Arg³⁸⁹ on TM6 by a salt bridge or hydrogen bond, respectively. The diphosphate group forms a salt bridge with Arg¹⁰⁴ on TM1 and Lys^336′ on TM5′, and a hydrogen bond with tyrosine (Tyr)^337′ on TM5′.⁵⁴⁾ In the ATP-bound structure the conformation changes to an outward-open state and the substrate is released. In addition, there is a helical break around proline (Pro)²⁶³ and Gly²⁶⁶ in TM4. A similar break in TM4 is found in the bacterial ABC transporter YbtPQ, which possesses the same TMD fold with ABCD transporter, and proposed as being key for substrate release.⁶¹⁾ The Arg²⁸⁰ existing at the cytosolic end of TM4 seems to seal the cavity from the peroxisomal lumen in the outward-open conformation. Arg²⁸⁰ is speculated to destabilize the outward-open conformation, resulting in an impairment of substrate transport.⁵⁶⁾ However, it is unclear whether acyl-CoA or free fatty acid binds to ABCD1 in outward-open state from the Cryo-EM pictures.⁵⁸⁾

Fig. 4. Schematic Illustration of the Deduced Conformational Changes of ABCD1 and ABCD4 during the Transport Cycle

(A) ABCD1 takes an inward-open conformation without substrate or ATP. The substrate-binding results in a narrower inward-open cavity. The presence of the substrate and ATP causes the transition from inward-open to the outward-open conformation, facilitating substrate release to the peroxisomal lumen. (B) ABCD4 takes a lysosomal-open conformation under the condition of ATP-binding to NBD. The hydrolysis of ATP triggers a conformation change of ABCD4 and release of cobalamin into the cytosol.

ABCD4 is thought to play an important role during the cobalamin transport from lysosomes to the cytosol. Recently, we demonstrated that ABCD4 possesses cobalamin transport activity across the liposomal membrane in vitro.²⁵⁾ ABCD4 had been purified according to the same procedure as ABCD1. We prepared ABCD4-liposomes including cobalamin and incubated them with the addition of ATP outside of the liposomes, and then quantified cobalamin release from the liposomes by reverse-phase HPLC. It became evident that ABCD4 is capable of transporting cobalamin from the inside to the outside of liposomes in a time-dependent manner. On the other hand, ABCD4(K427A), an ATPase activity-deficient mutant, had no cobalamin transport activity. Hence, ABCD4 transports cobalamin inside liposomes to the outside in an ATPase activity-dependent manner. As mentioned above, LMBD1 assists the translocation of ABCD4 from the ER to lysosomes through forming a complex with ABCD4. We examined whether LMBD1 has a different effect on ABCD4 function besides assisting in its translocation. However, LMBD1 itself has neither ATPase nor cobalamin transport activities. Subsequently, we prepared liposomes containing ABCD4 and LMBD1 as a complex. Liposomes containing the ABCD4-LMBD1 complex also showed both ATPase and cobalamin transport activity, but there was no significant difference in both activities compared with the ABCD4-liposomes. Thus, it is concluded that ABCD4 by itself is sufficient to transport cobalamin inside liposomes to the outside and that LMBD1 is not involved in cobalamin transport. It is predicted that the NBD of ABCD4 is face to the cytosol. Therefore, our system has accurately reproduced the cobalamin transport from the inside of lysosomes to the cytosol.

The Cryo-EM structure of human ABCD4 with ATP has been reported.⁶²⁾ The transmembrane cavity possesses an entrance constituted by hydrophobic residues followed by a narrowing cleft consisting of four pairs of negatively charged residues toward the cytosol, and finally gets closed at the membrane boundary. The hydrophobic entrance is similar to the substrate entrance at the TMDs of the Escherichia coli cobalamin ABC importer BtuCD. Indeed, an open hydrophobic cavity of BtuCD has been proposed to adapt the corrin ring moiety of cobalamin. Cobalamin probably binds to the TM cavity facing the lysosomes under conditions in which ATP binds to ABCD4. Subsequently, cobalamin is transported to cytosolic space through conformational changes in ABCD4 by ATP hydrolysis (Fig. 4B). Further study will provide insight into the transport mechanisms of ABCD1–4.

5. Human Diseases

Mutations of genes encoding ABCD1 and ABCD4 become causes of X-ALD and cobalamin deficiency, respectively. Disease-causing mutations and cryo-EM structures^54–58,62) have provided insight into the transport mechanism of ABCD1 and ABCD4.

Concerning ABCD1, at present, >900 disease-causing mutations have been reported in the X-ALD database (https://adrenoleukodystrophy.info/). The mutation types consist of missense (44%), frame shift (27%), nonsense (12%), amino acid insertion/deletion (5.5%), splice sites (4.7%), one or more exons deleted (2.5%) and benign variants (4.3%). Although 79% of all ABCD1 mutations affect the stability of ABCD1, the missense mutations, which have no impact on protein stability, indicate the residues of amino acids critical for function. Since the TMD forms the framework for the translocation pathway, we focused on TMD mutations. As mentioned above, there is a helical break around Pro²⁶³ and Gly²⁶⁶ in TM4 that is believed to be involved in substrate release.⁵⁶⁾ G266R is one of the most frequent mutations in X-ALD patients with a normal ABCD1 level, so Gly²⁶⁶ is deduced to be important for substrate release into the peroxisomal lumen. Arg²⁸⁰ is thought to be involved in the sealing of the cavity from the peroxisomal lumen in the outward-open conformation. R280C is found in 17 X-ALD patients with normal ABCD1 levels, and the mutation is speculated to destabilize the outward-open conformation, resulting in an impairment of substrate transport. Based on the substrate-bound structure, disease-causing mutations have been reported in Leu²²⁹, Trp³³⁹, Gly³⁴³ and Leu³⁹² in a cluster of hydrophobic residues crossing TM3–6 to which the acyl chain is bound.⁵⁴⁾ In particular, Gly³⁴³ is present in 25 cases. However, there are no data on ABCD1 stability for these residues. With regard to the pocket involved in the binding of the CoA moiety, as S213C, K217E, R401Q and R152C have been reported in patients with normal ABCD1 levels, these residues seem to be important for substrate binding, with R401Q being one of the most frequently reported missense mutants. R152L and R104C have also been identified in patients, but the ABCD1 level is reduced in these patients. It is therefore suspected that the reason is not the inability to bind substrate. As mentioned earlier, ABCD1 has been shown to transport free fatty acid after the hydrolysis of fatty acyl-CoA.^49,52) It has also been reported that Pxa1/Pxa2, the heterodimeric peroxisomal ABC transporter in S. cerevisiae, provides an import pathway for cleaved free CoA.⁵³⁾ Further analysis is needed to clarify the precise transport mechanism.

To date, eight clinical mutations in the ABCD4 gene that result in cobalamin deficiency have been reported. They include three flame shifts, two in-frame deletions, and three point mutations, N141K, Y319C, and R432Q. Arg⁴³² is located in the NBD, whereas asparagine (Asn)¹⁴¹ and Tyr³¹⁹ are located on the cytosolic side of TM3 and the lysosomal side of TM6, respectively.⁶²⁾ We examined a battery of ABCD4 missense mutations in an effort to elucidate how cobalamin transport is compromised by these alterations.²⁵⁾ We initially confirmed at least a part of both ABCD4(N141K) and ABCD4(Y319C) are located in lysosomes as well as the wild type. Subsequently, we evaluated cobalamin transport ability of ABCD4(N141K) and ABCD4(Y319C). ABCD4(N141K) has been deficient in cobalamin transport activity while retaining ATPase activity. From the cryo-EM structure, Asn¹⁴¹ faces the transmembrane cavity. The hydrophobic or ionic environment of the area might be important for cobalamin transport. Since cobalamin transport activity was restored in both the mutations of ABCD4(N141A) and ABCD4(N141D), an increase in the cationic charge around Asn¹⁴¹ by substitution to Lys¹⁴¹ might result in the impairment of cobalamin transport. On the other hand, ABCD4(Y319C) lacks both ATPase and cobalamin transport activity. It is speculated that Y319C mutant tends to form a disulfide bond via the Cys³¹⁹ residues of each monomer and lost cobalamin transport activity due to the impairment in the conformational change. However, as ABCD4(Y319A) also lost ATPase activity as well as cobalamin transport activity, the formation of disulfide bond is not the reason. The Y319C mutant has a further impairment in conformational change associated with dysfunction because ABCD4(Y319C) was able to bind with ATP without hydrolyzing it. Recently, the structural features of Cyanidioschyzon merolae ABCB1 during conformational realignment have been revealed based on high-resolution crystallographic analysis.⁶³⁾ The aromatic amino acid residues in the upper part of TM4 and TM6 create a hydrophobic cluster through van der Waals contacts and hydrogen-bonding networks. The amino acid residues are involved in the transition to outward-open configuration that take place along with ATP binding. The Tyr³¹⁹ in the TM6 of ABCD4 is deduced to be involved in the corresponding cluster and is important for the conformational change associated with ATP binding and hydrolysis. In fact, substituting Tyr³¹⁹ to Phe³¹⁹ with an aromatic ring did not result in a reduction of both ATPase and cobalamin transport activities. N141K and Y319C have an impact on enzyme activities, but there is no data on which residues are responsible for the binding with cobalamin.

6. Concluding Remarks

ABCD1–3 are localized on peroxisomal membranes and transport substrates from the cytosol to the peroxisomal lumen, whereas ABCD4 exists on lysosomal membranes and transports cobalamin from the lysosomal lumen to the cytosol. Although all ABCD transporters possess the same TMD architecture and topology, ABCD1–3 and ABCD4 transport substrates in opposite direction during conformational changes of the TMD associated with ATP binding and hydrolysis of the NBD. Recently, it was reported that some kind of ABC transporters in bacteria, which possess the same folding of TMD as the mammalian ABC exporters, function as importers.^61,64) Furthermore, there is a surprising report that the transport direction of P-glycoprotein (ABCB1) is altered from export to import by substituting a group of 14 conserved residues in homologous transmembrane helices 6 and 12 with alanine.⁶⁵⁾ Cryo-EM as well as X-ray crystallography has emerged as a powerful approach for determining the structure of proteins, and the cryo-EM structure of various ABC transporters has been determined. The underlying mechanism to determine the transport direction and substrate specificity of ABCD transporters will thus be elucidated in the near future. Furthermore, the physical interaction of the ABCD transporters with other proteins is important for understanding the role of ABCD transporters in various cellular processes in addition to the efficient regulation of transport.

Acknowledgments

This work was supported in part by Grants-in-Aid for Early-Career Scientists (18K14900, 20K15990) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and Tamura Science and Technology Foundation.

Conflict of Interest

The authors declare no conflict of interest.

References

• 1) Rees D. C., Johnson E., Lewinson O., Nat. Rev. Mol. Cell Biol., 10, 218–227 (2009).
• 2) Thomas C., Tampé R., Annu. Rev. Biochem., 89, 605–636 (2020).
• 3) Cui J., Davidson A. L., Essays Biochem., 50, 85–99 (2011).
• 4) Locher K. P., Nat. Struct. Mol. Biol., 23, 487–493 (2016).
• 5) Shitan N., Bazin I., Dan K., Obata K., Kigawa K., Ueda K., Sato F., Forestier C., Yazaki K., Proc. Natl. Acad. Sci. U.S.A., 100, 751–756 (2003).
• 6) Lee M., Choi Y., Burla B., Kim Y., Jeon B., Maeshima M., Yoo Y., Martinoia E., Lee Y., Nat. Cell Biol., 10, 1217–1223 (2008).
• 7) Quazi F., Lenevich S., Molday R. S., Nat. Commun., 3, 925 (2012).
• 8) Puljung M. C., J. Gen. Physiol., 150, 653–669 (2018).
• 9) Csanády L., Vergani P., Gadsby D. C., Physiol. Rev., 99, 707–738 (2019).
• 10) Vasiliou V., Vasiliou K., Nebert D. W., Hum. Genomics, 3, 281–290 (2009).
• 11) Gerovac M., Tampé R., Trends Biochem. Sci., 44, 167–180 (2019).
• 12) Fostier C. R., Monlezun L., Ousalem F., Singh S., Hunt J. F., Boël G., FEBS Lett., 595, 675–706 (2021).
• 13) Srikant S., Gaudet R., Nat. Struct. Mol. Biol., 26, 792–801 (2019).
• 14) Ueda K., Biosci. Biotechnol. Biochem., 75, 401–409 (2011).
• 15) Kamijo K., Taketani S., Yokota S., Osumi T., Hashimoto T., J. Biol. Chem., 265, 4534–4540 (1990).
• 16) Mosser J., Douar A. M., Sarde C. O., Kioschis P., Feil R., Moser H., Poustka A. M., Mandel J. L., Aubourg P., Nature (London), 361, 726–730 (1993).
• 17) Lombard-Platet G., Savary S., Sarde C. O., Mandel J. L., Chimini G., Proc. Natl. Acad. Sci. U.S.A., 93, 1265–1269 (1996).
• 18) Shani N., Jimenez-Sanchez G., Steel G., Dean M., Valle D., Hum. Mol. Genet., 6, 1925–1931 (1997).
• 19) Kawaguchi K., Morita M., Biomed. Res. Int., 2016, 6786245 (2016).
• 20) Geillon F., Gondcaille C., Charbonnier S., Van Roermund C. W., Lopez T. E., Dias A. M., Pais de Barros J. P., Arnould C., Wanders R. J., Trompier D., Savary S., J. Biol. Chem., 289, 24511–24520 (2014).
• 21) Tawbeh A., Gondcaille C., Trompier D., Savary S., Int. J. Mol. Sci., 22, 6093 (2021).
• 22) Ferdinandusse S., Jimenez-Sanchez G., Koster J., Denis S., Van Roermund C. W., Silva-Zolezzi I., Moser A. B., Visser W. F., Gulluoglu M., Durmaz O., Demirkol M., Waterham H. R., Gökcay G., Wanders R. J., Valle D., Hum. Mol. Genet., 24, 361–370 (2015).
• 23) Kawaguchi K., Okamoto T., Morita M., Imanaka T., Sci. Rep., 6, 30183 (2016).
• 24) Coelho D., Kim J. C., Miousse I. R., et al., Nat. Genet., 44, 1152–1155 (2012).
• 25) Kitai K., Kawaguchi K., Tomohiro T., Morita M., So T., Imanaka T., J. Biol. Chem., 296, 100654 (2021).
• 26) Kashiwayama Y., Seki M., Yasui A., Murasaki Y., Morita M., Yamashita Y., Sakaguchi M., Tanaka Y., Imanaka T., Exp. Cell Res., 315, 190–205 (2009).
• 27) Lee A., Asahina K., Okamoto T., Kawaguchi K., Kostsin D. G., Kashiwayama Y., Takanashi K., Yazaki K., Imanaka T., Morita M., Biochem. Biophys. Res. Commun., 453, 612–618 (2014).
• 28) Sacksteder K. A., Jones J. M., South S. T., Li X., Liu Y., Gould S. J., J. Cell Biol., 148, 931–944 (2000).
• 29) Biermanns M., Gartner J., Biochem. Biophys. Res. Commun., 285, 649–655 (2001).
• 30) Kashiwayama Y., Asahina K., Shibata H., Morita M., Muntau A. C., Roscher A. A., Wanders R. J., Shimozawa N., Sakaguchi M., Kato H., Imanaka T., Biochim. Biophys. Acta, 1746, 116–128 (2005).
• 31) Kashiwayama Y., Asahina K., Morita M., Imanaka T., J. Biol. Chem., 282, 33831–33844 (2007).
• 32) Sakaue H., Iwashita S., Yamashita Y., Kida Y., Sakaguchi M., J. Biochem., 159, 539–551 (2016).
• 33) Halbach A., Lorenzen S., Landgraf C., Volkmer-Engert R., Erdmann R., Rottensteiner H., J. Biol. Chem., 280, 21176–21182 (2005).
• 34) Cartier N., Lopez J., Moullier P., Rocchiccioli F., Rolland M. O., Jorge P., Mosser J., Mandel J. L., Bougnères P. F., Danos O., Proc. Natl. Acad. Sci. U.S.A., 92, 1674–1678 (1995).
• 35) van Roermund C. W., Visser W. F., Ijlst L., van Cruchten A., Boek M., Kulik W., Waterham H. R., Wanders R. J., FASEB J., 22, 4201–4208 (2008).
• 36) van Roermund C. W., Visser W. F., Ijlst L., Waterham H. R., Wanders R. J., Biochim. Biophys. Acta, 1811, 148–152 (2011).
• 37) Hama K., Fujiwara Y., Takashima S., Hayashi Y., Yamashita A., Shimozawa N., Yokoyama K., J. Lipid Res., 61, 523–536 (2020).
• 38) Pujol A., Ferrer I., Camps C., Metzger E., Hindelang C., Callizot N., Ruiz M., Pàmpols T., Giròs M., Mandel J. L., Hum. Mol. Genet., 13, 2997–3006 (2004).
• 39) Berger J., Albet S., Bentejac M., Netik A., Holzinger A., Roscher A. A., Bugaut M., Forss-Petter S., Eur. J. Biochem., 265, 719–727 (1999).
• 40) van Roermund C. W., Ijlst L., Wagemans T., Wanders R. J., Waterham H. R., Biochim. Biophys. Acta, 1841, 563–568 (2014).
• 41) Appelqvist H., Wäster P., Kågedal K., Öllinger K., J. Mol. Cell Biol., 5, 214–226 (2013).
• 42) Hunt A., Harrington D., Robinson S., BMJ, 349 (sep04 1), g5226 (2014).
• 43) Rutsch F., Gailus S., Miousse I. R., Suormala T., Sagne C., Toliat M. R., Nurnberg G., Wittkampf T., Buers I., Sharifi A., Stucki M., Becker C., Baumgartner M., Robenek H., Marquardt T., Hohne W., Gasnier B., Rosenblatt D. S., Fowler B., Nurnberg P., Nat. Genet., 41, 234–239 (2009).
• 44) Hillebrand M., Gersting S. W., Lotz-Havla A. S., Schäfer A., Rosewich H., Valerius O., Muntau A. C., Gärtner J., J. Biol. Chem., 287, 210–221 (2012).
• 45) Deme J. C., Hancock M. A., Xia X., Shintre C. A., Plesa M., Kim J. C., Carpenter E. P., Rosenblatt D. S., Coulton J. W., Mol. Membr. Biol., 31, 250–261 (2014).
• 46) Geillon F., Gondcaille C., Raas Q., Dias A. M. M., Pecqueur D., Truntzer C., Lucchi G., Ducoroy P., Falson P., Savary S., Trompier D., J. Biol. Chem., 292, 6965–6977 (2017).
• 47) Lotz-Havla A. S., Woidy M., Guder P., Friedel C. C., Klingbeil J. M., Bulau A. M., Schultze A., Dahmen I., Noll-Puchta H., Kemp S., Erdmann R., Zimmer R., Muntau A. C., Gersting S. W., J. Proteome Res., 20, 4366–4380 (2021).
• 48) Wiesinger C., Kunze M., Regelsberger G., Forss-Petter S., Berger J., J. Biol. Chem., 288, 19269–19279 (2013).
• 49) van Roermund C. W., Ijlst L., Majczak W., Waterham H. R., Folkerts H., Wanders R. J., Hellingwerf K. J., J. Biol. Chem., 287, 20144–20153 (2012).
• 50) De Marcos Lousa C., van Roermund C. W., Postis V. L., Dietrich D., Kerr I. D., Wanders R. J., Baldwin S. A., Baker A., Theodoulou F. L., Proc. Natl. Acad. Sci. U.S.A., 110, 1279–1284 (2013).
• 51) Okamoto T., Kawaguchi K., Watanabe S., Agustina R., Ikejima T., Ikeda K., Nakano M., Morita M., Imanaka T., Biochem. Biophys. Res. Commun., 496, 1122–1127 (2018).
• 52) Kawaguchi K., Mukai E., Watanabe S., Yamashita A., Morita M., So T., Imanaka T., Sci. Rep., 11, 2192 (2021).
• 53) van Roermund C. W. T., IJlst L., Baker A., Wanders R. J. A., Theodoulou F. L., Waterham H. R., FEBS Lett., 595, 763–772 (2021).
• 54) Chen Z. P., Xu D., Wang L., Zhou C. Z., Hou W. T., Chen Y., Nat. Commun., 13, 3299 (2022).
• 55) Jia Y., Zhang Y., Lei J., Yang G., bioRxiv:2021.2009.2024.461756.
• 56) Le L. T. M., Thompson J. R., Dang P. X., Bhandari J., Alam A., bioRxiv:2021.2009.2004.458904.
• 57) Xiong C., Jia L. N., Shen M. H., Xiong W. X., Xiong L. L., Wang T. H., Zhou D., Liu Z., Tang L., bioRxiv:2021.2010.2014.464310.
• 58) Wang R., Qin Y., Li X., Cell Res., 32, 214–217 (2022).
• 59) Mi W., Li Y., Yoon S. H., Ernst R. K., Walz T., Liao M., Nature (London), 549, 233–237 (2017).
• 60) Nosol K., Bang-Sørensen R., Irobalieva R. N., Erramilli S. K., Stieger B., Kossiakoff A. A., Locher K. P., Proc. Natl. Acad. Sci. U.S.A., 118, e2106702118 (2021).
• 61) Wang Z., Hu W., Zheng H., Sci. Adv., 6, eaay7997 (2020).
• 62) Xu D., Feng Z., Hou W. T., Jiang Y. L., Wang L., Sun L., Zhou C. Z., Chen Y., Cell Res., 29, 1039–1041 (2019).
• 63) Kodan A., Yamaguchi T., Nakatsu T., Matsuoka K., Kimura Y., Ueda K., Kato H., Nat. Commun., 10, 88 (2019).
• 64) Arnold F. M., Weber M. S., Gonda I., Gallenito M. J., Adenau S., Egloff P., Zimmermann I., Hutter C. A. J., Hürlimann L. M., Peters E. E., Piel J., Meloni G., Medalia O., Seeger M. A., Nature (London), 580, 413–417 (2020).
• 65) Sajid A., Lusvarghi S., Murakami M., Chufan E. E., Abel B., Gottesman M. M., Durell S. R., Ambudkar S. V., Proc. Natl. Acad. Sci. U.S.A., 117, 29609–29617 (2020).