Intricate Regulation of Sphingolipid Biosynthesis: An In‐Depth Look Into ORMDL‐Mediated Regulation of Serine Palmitoyltransferase

ABSTRACT

Sphingolipids are a structurally unique, widespread, and diverse family of lipids. Serine palmitoyltransferase (SPT) is the first and rate‐limiting enzyme required for the synthesis of all sphingolipids. Not unexpectedly, SPT is highly regulated. SPT is a multi‐subunit enzyme, the level of activity of which is controlled by the regulatory subunits known as the ORMDLs. Here, we discuss how the regulation of SPT activity is accomplished by multiple mechanisms, underscoring the importance of this regulation. A rapid homeostatic regulation of SPT, monitoring cellular sphingolipid levels, is mediated by the direct binding of the central sphingolipid ceramide to the SPT/ORMDL complex. This acute regulation is overlaid by a longer‐term regulation in which ORMDL is removed from the remainder of the SPT complex and trafficked for degradation, resulting in enhanced SPT activity. A third level of regulation is conferred by the inclusion of specific isoforms of the subunits of SPT into the complex. The isoform composition of the SPT complex dictates both the sensitivity of the complex to levels of cellular sphingolipid and the molecular species of sphingoid backbone that are produced. Here we discuss the mechanisms, interplay, and physiological roles of these three levels of regulation of sphingolipid biosynthesis.

Multiple enzymes play a crucial role in regulating the biosynthesis of de novo sphingolipids. This regulation starts with the rate‐limiting enzyme, serine palmitoyltransferase (SPT), which catalyzes the first step of the pathway. Disruptions in this regulatory process can lead to serious diseases. A variety of mechanisms are in place to control and shape SPT activity, ensuring that sphingolipid levels within the cell remain balanced.

Abbreviations

3‐KDS

3‐ketodihydrosphingosine

AOS

α‐oxoamine synthases

ATP

adenosine triphosphate

CerS

ceramide synthase

CERT

ceramide transfer protein

Co‐A

coenzyme A

Cryo‐EM

cryogenic electron microscopy

C‐terminal

carboxy‐terminal

DHS

dihydrosphingosine

ER

endoplasmic reticulum

HMG

3‐hydroxy‐3‐methylglutary

IC

inhibitory concentration

KDSR

3‐keto dihydrosphingosine reductase

N‐terminal

amino‐terminal

S1P

sphingosine‐1‐phosphate

SPT

serine palmitoyltransferase

TMD

transmembrane domain

1. Introduction

Sphingolipids form a major and diverse class of lipids. These unique lipids are essential constituents of cell membranes and serve as signaling molecules [1, 2]. An essential characteristic of sphingolipids is the sphingoid base backbone (Figure 1A) [3]. The sphingoid backbone can be modified and built on to generate the array of sphingolipid classes (Figure 1B,C). These include the sphingoid bases themselves, which may be phosphorylated to produce the potent signaling lipid sphingosine‐1‐phosphate [4]. Ceramides are produced by acylation of the nitrogen group. Fatty acids of chain lengths of 14 to greater than 26 carbons can be N‐acylated to the sphingoid backbone, producing a large variety of ceramide species [5]. The fatty acid is appended by one of the six ceramide synthases, each of which has specificity for the length of the fatty acyl‐CoA substrate, thus determining the range of ceramide species generated. Ceramide can be further modified at the 1’hydroxyl. The addition of phosphorylcholine at the 1’hydroxyl generates sphingomyelin, a major plasma membrane lipid [6]. The addition of carbohydrate to the 1’ hydroxyl generates a range of glycosphingolipids, which have roles in intercellular signaling [7]. The 1’ hydroxyl can also be phosphorylated to produce ceramide‐1‐phosphate, a signaling lipid implicated in the generation of eicosanoids [8].

FIGURE 1.

Classification of sphingolipids. (A) Structure of Sphingosine. Sphingosine is the backbone of all sphingolipids. The hydrophilic head of the molecule has the capacity to form hydrogen bonds, a key feature of all sphingolipids. The 1’ hydroxyl can be modified, and the nitrogen can be acylated to generate the more complex sphingolipids. (B) The range of sphingolipids built on the sphingosine base. The sphingolipid family of lipids can be subdivided into sphingolipids that include the sphingoid bases and ceramide, and the complex lipids based on modifications of ceramide 1’ hydroxyl. (C) Structures of major sphingolipids built on the sphingosine backbone. Shown are ceramide, resulting from fatty acid addition to the sphingosine nitrogen, sphingomyelin, resulting from addition of phosphorylcholine to the 1’ hydroxyl of ceramide, and glucosyl‐ceramide, resulting from addition of glucose to the 1’ hydroxyl of ceramide.

Each class of sphingolipid plays specific roles in the cell. For example, sphingosine‐1‐phosphate is a well‐established signaling lipid with both extracellular and intracellular functions [4]. Ceramide serves a signaling function, best characterized by its role in the initiation of apoptosis, and acts as a central metabolite for the generation of other, more complex, sphingolipids. Sphingomyelin plays an essential role in establishing the physical properties of the plasma membrane [9]. Glycosphingolipids are essential for maintaining cell membrane integrity and are involved in cellular functions such as cell recognition, adhesion, and signaling [7].

Dysregulation of sphingolipid metabolism has been linked to various diseases, including neurodegenerative disorders, cancer, and metabolic disorders [10, 11, 12]. Particularly germane to this review, two classes of mutations in the serine palmitoyltransferase (SPT) enzyme complex, the initiating enzyme in the sphingolipid biosynthetic pathway (discussed in more detail below), are the basis of devastating neurological diseases (reviewed in [13]).

As the initiating and rate‐limiting enzyme of the de novo sphingolipid biosynthesis pathway, SPT is highly regulated. Considering the importance of sphingolipids and the profound consequences of dysregulation of sphingolipid metabolism, it is essential to understand the molecular mechanisms regulating cellular sphingolipid levels. There are multiple branches in the sphingolipid metabolic network that can affect the levels of individual sphingolipid species. However, here we focus on the regulation of SPT, which governs the overall levels of cellular sphingolipids.

2. The Sphingolipid Biosynthesis Pathway

Our review focuses on the initiating step in sphingolipid biosynthesis, but here we very briefly review the sphingolipid metabolic network for clarity. A number of excellent reviews focus on other steps in the overall metabolic pathway [2, 14, 15].

De novo sphingolipid biosynthesis is a multi‐step process that involves the coordination of several enzymes to produce a range of sphingolipids. It begins in the endoplasmic reticulum (ER) with the condensation of an amino acid (usually L‐serine) and a fatty acyl Co‐A (usually the 16‐carbon fatty acid palmitate) by the first and rate‐limiting enzyme of the de novo sphingolipid pathway SPT (Figure 2). The regulation of this enzyme is the focus of this review and is considered in more detail below. SPT's condensation reaction produces 3‐ketodihydrosphingosine (3‐KDS) [16]. 3‐KDS, in the next step, is reduced to dihydrosphingosine (DHS) by 3‐keto dihydrosphingosine reductase (KDSR) [16].

FIGURE 2.

Overview of de novo sphingolipid biosynthesis pathway in mammals. The initiating and rate‐limiting step in the production of all sphingolipids is catalyzed by the enzyme Serine Palmitoyltransferase (SPT). Subsequent reactions produce the key intermediate, ceramide. These steps all occur in the endoplasmic reticulum (ER). Ceramide may then be transported to the Golgi apparatus. The ceramide transporter CERT transfers ceramide from the ER to the trans Golgi, where ceramide is modified by phosphorylcholine, to produce sphingomyelin. Vesicular trafficking transfers ceramide to the Golgi where sugars, either glucose or galactose, are added to produce the precursor of more complex glycosphingolipids. Critically, all of the reactions subsequent to the production of ceramide are reversible by degradative enzymes. These include ceramidases, which generate sphingosine and a fatty acid. The released sphingosine can be phosphorylated by one of the two sphingosine kinases to produce sphingosine‐1‐phosphate (S1P), a potent signaling lipid and the substrate for S1P lyase, the only enzyme in this pathway capable of degrading the sphingosine backbone. The sphingomyelinases and glycosidases liberate ceramide (and phosphorylcholine or free sugars, respectively) that can then undergo additional metabolism. Blue = enzymes catalyzing each step. Green = non‐sphingolipid substrates and products of reactions.

Dihydrosphingosine is N‐acylated by the ceramide synthases. In mammals, there are six isoforms of this enzyme (Cers 1–6) and each has specificity for the chain‐length of fatty acid added to the sphingoid backbone [5]. Typically, in mammalian species the fatty acids added can range from 14 to 26 carbons. The desaturation of these dihydroceramides at carbons 4/5 by dihydroceramide Δ4‐desaturase 1 produces ceramide, the central metabolite of the sphingolipid biosynthesis pathway [17].

In mammalian cells, the production of higher‐order sphingolipids requires ceramide to be transported to the Golgi apparatus. At the medial Golgi, sphingomyelin is produced from ceramide by sphingomyelin synthase 1 [6, 18]. The ER to medial Golgi transport of ceramide for sphingomyelin production is accomplished by the action of ceramide transfer protein (CERT) [19]. Alternately, ceramide may be converted to glucosylceramide or galactosylceramide in the trans‐Golgi. Transport from the ER, in this case, is mediated by vesicular transport [20]. These glycolipids serve as the basis for the synthesis of a large series of more complex glycosphingolipids (reviewed in [7]). Ceramide may also be phosphorylated by ceramide kinase to produce the signaling lipid ceramide‐1‐phosphate (reviewed in [21]).

It is notable that in plants, glycosphingolipid synthesis occurs in the endoplasmic reticulum [22], while in insects, a sphingomyelin synthase‐related protein is found in the endoplasmic reticulum (reviewed in [6]). In mammalian cells the regulation of higher‐order sphingolipids can be controlled by ER to Golgi transport mechanisms, while this does not appear to be the case in plants and insects.

An interesting feature of the sphingolipid metabolic pathway is that once the sphingoid backbone is produced, it can circulate among all of the sphingolipid species through a series of both synthetic and degradative steps, with the backbone itself remaining intact. For example, ceramide may be hydrolyzed by one of several ceramidases, removing the fatty acid chain to generate sphingosine [23]. The sphingosine produced by these enzymes can be converted back into ceramide by the ceramide synthases by what is known as the salvage pathway [24]. Alternately, sphingosine may be phosphorylated to produce the potent signaling molecule sphingosine‐1‐phosphate (S1P) by one of two sphingosine kinases [4]. Sphingosine‐1‐phosphate can be reconverted back into sphingosine by the action of sphingosine‐1‐phosphate (S1P) phosphatases [25]. Similarly, sphingomyelin can be hydrolyzed by one of several sphingomyelinases to generate ceramide, which can then be further hydrolyzed to generate sphingosine or re‐utilized by sphingomyelin synthase [26]. Glycosphingolipids undergo hydrolysis to ceramide, principally in the lysosome. Defects in this hydrolysis are the basis of debilitating neurological diseases [27]. Thus, once produced, the sphingoid backbone can encounter numerous and dynamic, fates. However, there is only one step that leads to the ultimate degradation of that backbone, and this requires the conversion of sphingosine into sphingosine‐1‐phosphate. Sphingosine‐1‐phosphate can then be degraded into ethanolamine phosphate and hexadecenal by sphingosine‐1‐phosphate (S1P) lyase [28]. This is a critical step for controlling cellular levels of sphingolipid, and defects in it have profound effects at the cellular and organismal level [28] (Figure 2).

Thus, the sphingolipid metabolic network is unique in the sense that sphingolipid synthesis is dependent on a single reaction and can, ultimately, only be degraded through another. Here, we focus on the initiating and rate‐limiting step for the generation of the sphingoid backbone, the SPT enzyme complex, and how this enzyme is regulated to control cellular levels of sphingolipids. The bulk of our discussion is on the mammalian form of SPT.

3. The Architecture of the SPT/ORMDL Complex

3.1. The Catalytic Core of SPT

SPT, is a member of the α‐oxoamine synthase (AOS) family of enzymes [29]. The members of this enzyme family catalyze condensation reactions between acyl‐CoA thioesters and specific amino acid substrates [29]. SPT is a multi‐subunit enzyme complex bound to the ER membrane [30, 31]. In mammals the complex consists of four subunits, two of which form the catalytic core, one which draws the complex close to the ER membrane, and a fourth which is a regulatory subunit. The catalytic site of the enzyme is formed from two major subunits in most cells and tissues, SPTLC1 and SPTLC2. Both subunits are involved in the formation of the catalytic site, with SPTLC2 containing the covalently bound pyridoxyl phosphate co‐factor. Although SPTLC2 is most commonly found in the complex, in some tissues, and under some conditions, it can be replaced by an alternate isoform, SPTLC3. The consequence of this replacement on acyl‐CoA substrate specificity is outlined below. A third subunit is the so‐called small subunit, of which there are two isoforms, SPTssa and SPTssb [32]. The small subunit boosts the activity of the SPT complex by approximately 100‐fold [32], possibly by drawing the complex closer to the ER membrane to enhance both access to the hydrophobic acyl‐CoA substrate and the egress of the 3‐keto sphingosine product [30, 31] although other, structural changes promoting enzyme activity are also possible.

The specific isoform composition of the catalytic core of SPT determines the carbon chain length of the resulting sphingoid backbone [32]. The combination most commonly observed, SPTLC1/SPTLC2/SPTssa, prefers palmitoyl CoA as the acyl‐CoA substrate and thus produces an 18 carbon sphingoid base. The SPTLC1/SPTLC3/SPTssa complex can accommodate shorter acyl‐CoAs (12 and 14 carbon) and so produces 14 and 16 carbon bases [33]. The inclusion of SPTssb allows acyl‐CoAs both shorter and longer than 16 carbons, so when it is included in the complex, SPT produces sphingoid bases with an array of chain lengths [32]. As discussed below, it is critical to recognize that conditions that impact the expression of individual isoforms of SPT core subunits have a profound role in shaping the qualitative composition of cellular sphingolipids.

3.2. The Regulatory Subunits of SPT: The ORMDLs

The fourth subunit of SPT is ORMDL, of which there are three isoforms (ORMDL1‐3) in mammalian cells and tissues. It has been recognized for decades that SPT is homeostatically regulated; SPT activity is diminished when cellular sphingolipid levels are elevated [34]. Molecular genetics in yeast revealed that this regulation was dependent on the Orms (of which there are two isoforms), yeast homologues of the ORMDLs [35, 36]. Subsequently, the ORMDLs were established to fulfill the same function in higher eukaryotes [35, 37]. The ORMDLs are intrinsic subunits of the SPT complex, as determined both by immunoprecipitation studies [35, 37] and, as detailed below, by structural determination. There is a common reference to the ORMDLs as inhibitors of SPT. A more accurate view is that the ORMDLs are regulatory subunits of the SPT complex, which allow this enzyme to respond to cellular levels of sphingolipids. The remainder of this review addresses the mechanisms by which the ORMDLs regulate overall SPT activity and, in doing so, establish overall sphingolipid levels. Earlier reviews have addressed some of the basic concepts of ORMDL‐dependent regulation of SPT [38, 39]. However, very recent structural, cellular, and biochemical discoveries have painted a more complete and complex picture of this regulation.

4. Structure of the SPT Complex

Cryogenic Electron Microscopy (Cryo‐EM) was used to determine the first structures of the human SPT complex in 2021 by groups led by Xin Gong of Southwestern Institute of Science and Technology in China and Chia‐Hsueh Lee of St. Jude's Children's Research Hospital [30, 31]. Details of these structures, including the substrate binding sites and catalytic center, have been recently reviewed, and will not be covered in depth here [40]. Structures have recently been determined for the SPT complex in yeast and plants [41, 42, 43, 44]. Among these disparate species, the structures are remarkably similar, with two notable exceptions. The human and Arabidopsis forms of the enzyme are homodimers, with each monomer containing all four subunits. The yeast enzyme exists both in monomeric and dimeric forms [41, 42, 43]. Moreover, the yeast Orms are extended at the amino terminus by almost 70 amino acids relative to the mammalian ORMDLs. In this review, we focus on the human SPT complex.

The structure of the human SPT complex was determined by both groups by analyzing the SPT complex containing SPTLC1, SPTLC2, SPTssa, and ORMDL3. As illustrated in Figure 3, the SPT/ORMDL complex appears as a symmetrical dimer, with each monomer (monomer A and monomer B) consisting of SPTLC1, SPTLC2, SPTssa, and ORMDL3. The SPT complex comprises six transmembrane domains (TMD), with SPTLC1 and SPTssa contributing one TMD each, while ORMDL has four TMDs [30, 31]. In addition, SPTLC2 contains an amphipathic helix that appears to interact with the membrane bilayer [30, 31].

FIGURE 3.

Cryo‐EM structure of SPT ORMDL complex (PDB ID: 7k0n). The SPT‐ORMDL complex is a dimer (shown here as monomer A and monomer B) protein assembly with each monomer consisting of four different subunits. Depicted here is the structure determined by cryo‐EM microscopy [30, 31, 45]. There are two major subunits (SPTLC1 shown in blue) and SPTLC2 (shown in pink) that comprise the catalytic center. The small subunit SPTssa (shown in green) accelerates catalysis and specifies the acyl‐CoA substrate. ORMDL3 (shown in yellow) is the regulatory subunit of the complex, controlling activity of the enzyme in response to cellular sphingolipid levels. The structure shown is based on PDB ID: 7k0n. Notably, there are alternate subunits for SPTLC2 (SPTLC3), SPTssa (SPTssb), and ORMDL3 (ORMDLs 1 and 2). Structures with these alternate subunits have yet to be determined.

The structural basis for the shift in acyl‐CoA substrate specificity, dependent on which small subunit isoform is included in the complex, was also revealed by these structures [30, 31]. The substrate binding tunnel of SPTLC2 is occupied by the side chain of methionine 28 (Met28) of SPTssa. In the structure of the 3‐ketosphingosine (the product of the enzyme) bound SPT complex, Met28 is in close proximity to the acyl tail. Acyl‐CoA substrates longer than 16 carbons would experience a steric clash. In contrast, ssSPTb has a valine (Val28) in the place of Met28. Val28 has a smaller side chain, which would allow it to accommodate the longer chain acyl‐CoA substrates [30, 31].

The amino‐terminal transmembrane helix of SPTLC1 interacts with the TMD 4 of ORMDL3, while the cytosolic side amino acid residues of ORMDL3 interact with the amino acid residues of both SPTLC1 and SPTLC2. Additionally, the C‐terminal amino acid residues of SPTssa interact with the amino acid residues in luminal loop 1 of ORMDL3 [30, 31].

5. ORMDL‐Dependent SPT Regulation‐1991‐2022

In 1991, Konrad Sandhoff and colleagues demonstrated that the addition of sphingosine to cultured cells resulted in reduced SPT activity [34], a classic homeostatic regulation. The molecular basis of this phenomenon was established in 2010, as noted above, when two different groups, headed by Jonathon Weissman and Amy Chang, respectively, identified the yeast ORMs as essential for the homeostatic regulation of SPT [35, 36] and Weissman's group demonstrated that the ORMDLs serve a similar function in mammalian cells. Building on this work, in 2012, our team directly demonstrated that the addition of soluble, short‐chain ceramides in cells resulted in an ORMDL‐dependent inhibition of SPT activity and that the ORMDL isoforms were functionally redundant. These experiments provided compelling evidence that ORMDLs are required to sense sphingolipid levels in cells and regulate SPT activity [37].

Critically, structures reported by the Chia‐Hsueh group, and subsequently by Gong's group, pointed to how ORMDLs inhibit SPT [31, 45]. In these structures, in which the substrate binding site is unoccupied by either substrate or substrate analogues, the amino‐terminus of ORMDL occupies the substrate binding pocket. This strongly suggests that SPT is regulated by the presence or absence of the ORMDL amino‐terminus in the catalytic site. The functional test of this occupancy in cells and the isolated complex demonstrated that deletions of the amino‐terminus of ORMDL block its ability to inhibit SPT [45]. Together, the structural and functional studies establish that the ORMDL‐dependent inhibition of SPT is accomplished by the amino‐terminus of ORMDL blocking substrate binding.

However, two major questions remained to be answered. First, what is the sphingolipid(s) that is being monitored by this system, and secondly, how do increases in levels of this sphingolipid act through ORMDL to inhibit SPT? A preliminary answer to the first of these questions was provided by the observation that although sphingosine treatment results in inhibition of SPT, that inhibition is eliminated if the conversion of sphingosine to ceramide is blocked by a specific inhibitor of the ceramide synthases, Fumonisin B1 [46]. This demonstrated that the sphingolipid sensed by this system is ceramide or a downstream metabolite of ceramide.

A significant advance was the discovery by our group that the ORMDL‐dependent inhibition of SPT by ceramide could be reconstituted in a cell‐free system consisting of microsomal membranes [46]. This inhibition was independent of cytosolic factors and ATP, conditions that are incompatible with virtually all post‐translational modifications. There were no changes in ORMDL protein levels under these conditions. This finding demonstrated that ceramide was directly impacting SPT activity and that this was an ORMDL‐dependent mechanism. This contrasts with one of the known mechanisms of regulation of SPT activity by the yeast Orms, in which phosphorylation of the amino‐terminal extension of the yeast Orms by the TORC2‐dependent enzyme Ypk1 plays a role [35, 47]. The cell‐free system was useful in further defining the sphingolipids that regulate SPT. Neither sphingomyelin nor sphingosine were active and the potency of ceramide was directly related to the length of the N‐acylated fatty acid. The longer the chain, the more potent the ceramide. A pivotal observation was that the response was stereospecific with respect to ceramide. Ceramide contains two chiral centers, generating four potential stereoisomers, of which D‐erythro ceramide is the only native stereoisomer. Remarkably, it was only the D‐erythro stereoisomer that triggered inhibition of SPT. This was a strong indication that there was a specific binding site for ceramide that accomplished SPT inhibition, rather than a role for the well‐documented ability of ceramide to influence the physical properties of biological membranes [48]. Because these experiments were performed with native membranes, the identity of the protein containing the site remained unclear. It was attractive to speculate that it was within the SPT complex itself, perhaps in ORMDL, but in principle another protein could contain that site.

6. ORMDL‐Dependent SPT Regulation‐2023 and Onward: Identification of a Regulatory Ceramide Binding Site in SPT

Confirmation that there was a specific protein binding site within SPT for ceramide that induces inhibition of SPT, and identification of that site, was achieved in 2023 by the Gong group with experimental collaboration with our group [45]. Remarkable high‐resolution SPT structures identified a molecule of ceramide bound at the interface between ORMDL transmembrane helices 1 and 2 and SPTLC2 (Figure 4). Four amino‐acid residues of ORMDL3 and SPTLC2 are predicted to form hydrogen bonds with the polar head group of ceramides. These residues include side chain groups of Asn13 and His85 in ORMDL3, Tyr122 in SPTLC2, and the main chain group of Phe63 in ORMDL3 [45]. A virtually identical binding site was predicted by computational docking of ceramide to the SPT complex (Figure 4A,B) [49]. Proof that the identified ceramide binding site was functional emerged from functional tests of mutations in Asn13 of ORMDL3 and Tyr122 of SPTLC2. Both of these mutations dramatically reduced the response of SPT to ceramide.

FIGURE 4.

Ceramide binds to the SPT/ORMDL complex to regulate SPT activity. (A) The site of ceramide bound to the SPT/ORMDL complex. Ceramide binding, as determined by molecular docking [49], is shown in black. Also shown is the amino‐terminus of ORMDL (wheat orange spheres) occupying the SPT substrate‐binding site in the inhibitory conformation (based on PDB:7YIU). (B) Critical hydrogen bonds between ceramide (black) and ORMDL3 (yellow) and SPTLC2 (pink). A magnified view of the ceramide‐binding site, both predicted by molecular docking and determined by Cryo‐EM [45, 49], predict hydrogen bonding between the headgroup of ceramide and Asn13 of ORMDL3, and Tyr122 and Thr82 of SPTLC2. (C, D) A model of the ORMDL‐mediated, ceramide‐sensitive regulation of SPT. (C) When sphingolipid levels are low enzyme substrates have access to the active site (arrows indicate the substrate binding site and location of the essential catalytic cofactor 5’ pyridoxyl phosphate) and SPT is active. Note that this structure is derived from a ceramide‐free form of the complex accomplished by mutation of the critical ceramide binding residue (Asn13) (PDB:7YJ2). In this structure, the amino‐terminus of ORMDL3 is not resolved, indicating it is highly flexible. Asn11 (wheat orange sphere) is the first resolved residue of ORMDL3. (D) Elevation of cellular ceramide (black spheres) leads to ceramide binding to the SPT complex, stabilizing the amino terminus of ORMDL3 (wheat orange spheres) in the substrate‐binding pocket, inhibiting SPT activity.

Thus, beginning from a prediction in an in vitro system to the revolutionary power of cryo electron microscopy in determining high‐resolution structures of membrane proteins, we now understand at a detailed level how cells sense sphingolipid levels to control sphingolipid production. A model depicting this regulation is illustrated in (Figure 4C,D). There remain some detailed mechanistic questions. From the structures alone, it is not clear how ceramide binding stabilizes the amino‐terminus of ORMDL in the mammalian SPT substrate binding site or the inhibitory Orm conformations in other species. Additionally, the influence of the membrane bilayer on catalytic aspects of SPT and on the ORMDL‐dependent regulation remains to be clarified. All of the structures determined to date are of detergent‐solubilized complexes. Structures from the SPT complex reconstituted into lipid nano discs should give a picture of the complex in its native environment. Molecular dynamic simulations could help elucidate a number of issues, including the trajectory of ceramide entering from the membrane bilayer into the regulatory binding site, how that binding stabilizes the amino‐terminus of ORMDL in the SPT catalytic site, and how substrates and products enter and egress, respectively, from the catalytic site.

In the past year structures of the Arabidopsis and yeast SPT complexes also resolved ceramide bound to the complex and, amazingly, the binding site is in essentially the same position as it is in the mammalian complex [41, 43]. For yeast this complements biochemical analysis that demonstrated that the direct regulation of yeast SPT by ceramide overlays the previously established regulation by phosphorylation of the N‐terminal extension of Orm (relative to ORMDL) by Ypk1, mentioned above [47]. A comparison of the similarities and differences of ceramide binding and the mode of inhibition by the ORMDL/Orm amino terminus has been recently reviewed and so will not be addressed in detail here [49]. Suffice it to say that the mode by which the amino‐terminus of ORMDL regulates SPT activity differs significantly in all three species. Rather than directly inserting into the substrate binding site, the Arabidopsis Orm amino‐terminus lies more superficially at the presumptive entrance to that site, blocking entry. With ceramide bound, the yeast Orm amino terminus is situated completely outside of the catalytic site and is presumed to have an allosteric effect on enzyme activity. The conservation of ceramide binding to the complex of these disparate species speaks to the fundamental importance cells place on regulating ceramide levels. It will be fascinating to uncover the evolutionary and functional basis that drives ceramide binding to use different modes of SPT regulation by the ORMDL/Orm amino terminus.

7. Layers Upon Layers of SPT Regulation

The direct binding of ceramide to the SPT complex is a straightforward and acute mechanism of homeostatic regulation. Not surprisingly, however, given the critical position that SPT plays in sphingolipid metabolism, there are additional layers of regulation that have come to light. As reviewed below, some of these mechanisms change the qualitative catalytic properties of the enzyme, while others impact the overall level of enzyme activity (Figure 5A–C).

FIGURE 5.

Modes of SPT regulation. (A) Direct binding of ceramide: Ceramide binding to the SPT/ORMDL complex leads to inhibition of SPT as outlined in Figure 4. (B) ORMDL degradation‐based SPT regulation: Under conditions of cholesterol overload [58] or reduced S1P levels ORMDL is selectively removed from the SPT complex and degraded, resulting in upregulation of SPT activity. (C) Subunit composition of the SPT complex. Aspects of SPT activity are affected by the specific subunits of SPTLC2/3, SPTssa/b, and ORMDL1‐3 incorporated into the complex. Inclusion of SPTLC3 in the place of SPTLC2 promotes the use of a 14 carbon acyl‐CoA by SPT rather than the 16 carbon acyl‐CoA used when SPTLC2 is incorporated. The small subunits, SPTssa or SPTssb, both enhance activity of the complex, however inclusion of SPTssb expands acyl‐CoA preference beyond palmitoyl CoA. Each ORMDL subunit imparts a distinctive level of ceramide sensitivity to the complex, with ORMDL3 being the most sensitive and ORMDL2 being the least. In principle, as shown, this affects the steady‐state levels of sphingolipid. Not shown is the regulator Nogo B, which negatively regulates SPT [60].

7.1. Regulation of SPT Acyl‐CoA Specificity by Subunit Isoform Selection

As noted above, incorporation of SPTssa versus SPTssb into the SPT complex impacts the acyl‐CoA substrate selection of SPT and therefore the length of the hydrocarbon tail of the resulting sphingosine backbone. The inclusion of SPTssa dictates that only 16‐carbon or shorter acyl‐CoAs can be utilized, while the inclusion of SPTssb allows longer substrates to be used. The incorporation of SPTLC3 in place of SPTLC2 promotes a preference for acyl‐CoAs of 12 and 14 carbons and, therefore, further amplifies the range of acyl‐CoA substrates used and the subsequent structure of the sphingoid backbone [33]. Expression levels of these alternate subunits, and the changes in sphingoid backbone that result, have physiological consequences. As an example, SPTLC3 expression is induced in ischemic cardiomyopathy both in humans and mouse models [50]. Lipidomic analysis reveals the expected accompanying increase in d16 sphingoid bases, derived from the SPTLC3‐driven utilization of 14‐carbon acyl‐CoAs by SPT. Depletion of SPTLC3 in a mouse model attenuates the effects of ischemia, demonstrating the functional consequences of SPTLC3 incorporation into the SPT complex [50]. This pathophysiological effect of SPTLC3 expression in this instance, no doubt, points to a as yet identified physiological role for the incorporation of SPTLC3 into the complex.

Skin ceramides are by far the most complex known in any tissue. They include a variety of sphingoid bases and N‐linked fatty acids not found elsewhere in similar abundance in the body [51]. The expression levels of all subunits of SPT increase substantially during skin keratinocyte differentiation. But the increased expression of SPTLC3 and SPTssb is especially dramatic [51]. Functional tests of these increases have yet to be performed, but it would hardly be surprising to find that the skin permeability barrier is significantly dependent on the variety of sphingoid bases produced by the inclusion of SPTLC3 and SPTssb during differentiation of skin keratinocytes.

The myelin membrane is a lipid‐rich membrane that enwraps axons in the central and peripheral nervous systems to accelerate nerve conduction and support neuronal health (reviewed in [52]. This unique membrane is especially rich in sphingolipids. These lipids are critical for the formation and function of myelin. The level of SPTLC3 is initially high in oligodendrocytes, the myelin‐producing cells of the central nervous system [53]. These levels decline as myelination proceeds. Conversely, expression levels of SPTssb increase dramatically and remain high. This parallels the hydrocarbon chain length of the sphingoid bases in oligodendrocytes. The content of d16 sphingolipids (derived from 14‐carbon acyl‐CoA, preferred by SPTLC3) declines slightly as myelination reaches its peak, while the content of d20 sphingolipids (derived from 18‐carbon acyl‐CoA, preferred by SPTssb) increases dramatically [53].

These three examples demonstrate that there is a regulation of SPT activity driven by the incorporation of alternate subunits that, by changing the nature of the sphingoid base produced, has important physiological roles in maintaining the function of specialized biological membranes. Many more examples of this mode of SPT regulation are sure to emerge.

7.2. Regulation of SPT Sensitivity to Ceramide by ORMDL Isoform Expression

Why are there three ORMDL isoforms? In culture the isoforms are redundant. Expression of any single isoform regulates the response of SPT to ceramide to the same extent as expression of all three together [37]. Even in the intact animal, the deletion of two of the three ORMDLs (ORMDL1 and ORMDL3) has relatively mild effects, expressed as defects in myelination, but otherwise has no other overt phenotype [54]. It is worth noting, however, that deletion of all three ORMDLs is embryonic lethal [54]. One rather mundane possibility is that distinct sequences in the promoters of the three isoforms allows specificity to the control of overall ORMDL content in different tissues and under different conditions. Although this may be true, recent evidence suggests that depending upon which isoform(s) is expressed, cells are tuned to respond to different levels of the ceramide pool. Xin Gong's group has measured the dose‐response to ceramide of isolated SPT complexes containing each of the ORMDL isoforms [45]. Our laboratory has made similar comparisons in membranes isolated from cells, engineered by Crispr/Cas9 knockout, expressing only one each of the ORMDL isoforms [55]. The results from these two systems agree. SPT complexes containing ORMDL3 are the most sensitive to ceramide levels, followed by ORMDL1 and then ORMDL2. In our hands, based on a measurement of IC₅₀, ORMDL3 is approximately three‐fold more sensitive than ORMDL1, and ORMDL1 is approximately 1.5 more sensitive than ORMDL2. The measurements in the isolated complex by Dr. Gong's group demonstrate more dramatic differences, but in the same qualitative order. Studies in yeast also point to Orm isoform specificity in terms of SPT inhibitory capacity. Frohlich and colleagues have demonstrated that Orm2 has a more strongly inhibitory effect on SPT than Orm1 [41]. More work needs to be done in this area, but the hints are there. For example, during myelination the expression levels of ORMDL1 and 3 decline, while levels of ORMDL2 remain constant [53]. This may be a mechanism to reduce feedback inhibition by ceramide to enhance SPT activity during a period when sphingolipid synthesis for the production of myelin is ramped up. A note of caution is required at this point regarding determining the protein levels of the individual ORMDL isoforms. Currently, relative levels of individual ORMDLs are measured at the mRNA level. Protein levels may be different. No antibody can distinguish the individual isoforms. The amino acid sequence differences between the three isoforms are so dispersed that the production of isoform‐specific antibodies is virtually impossible. A solution to this issue will likely involve the production of cell or animal lines with knocked‐in epitope tags.

7.3. Regulated Degradation of ORMDL

The direct binding of ceramide to the SPT/ORMDL complex can provide short‐term regulation of SPT. However, additional mechanisms are in play in which the ORMDLs are selectively removed from the SPT/ORMDL complex and subsequently degraded, eliminating ORMDL regulation of SPT activity.

This mechanism appears to be evolutionarily conserved. In yeast, Orm2 is phosphorylated upon reduced cellular sphingolipid levels, transported to the Golgi apparatus and endosomes, and then degraded in a process termed Endosome and Golgi‐Associated Degradation (EGAD) [56, 57]. The consequence of this degradation, to counter the reduced sphingolipid content in the cells, is reduced Orm2 inhibition of SPT and therefore increased sphingolipid production. The first report of this regulated degradation in mammalian cells was described in cholesterol‐loaded macrophages by Kailash Gulshan and colleagues [58]. This group was interested in the mechanism underlying increased sphingomyelin levels in atherosclerotic lesions and thought that increased SPT activity might be the underlying cause. Indeed, they found that when high levels of cholesterol were induced in a macrophage cell line SPT activity was increased, and that this was accompanied by a progressive loss of ORMDL content over several hours. They went on to demonstrate that the half‐life of ORMDL decreased from 11 to approximately 3 h under these conditions. This increased degradation is mediated by autophagy. Teleologically, the increased SPT activity with cholesterol loading makes sense to balance the well‐established relationship between cholesterol and sphingomyelin in biological membranes [59]. What remains to be established is exactly how cholesterol stimulates ORMDL degradation and how ORMDL is extracted from the SPT complex and incorporated into autophagic vesicles.

Vascular endothelial cells are exquisitely responsive to sphingolipids, in particular ceramide and sphingosine‐1‐phosphate (S1P) [60]. This has apparently led to unique mechanisms by these cells control sphingolipid synthesis. Di Lorenzo and colleagues have established that extracellular S1P engagement with one of its cell surface receptors (S1PR1) results in the stabilization of ORMDLs [61]. Conversely, when S1P levels decline or when signaling through the S1P receptor is blocked, ORMDLs undergo accelerated degradation, resulting in SPT complexes lacking ORMDL control. The consequence is higher SPT activity, presumably to maintain S1P levels. In this case enhanced ORMDL degradation is mediated by proteasomal activity, in contrast to the autophagic degradation resulting from increased cholesterol levels. It remains to be seen which other cell types utilize S1P‐mediated ORMDL stabilization/degradation as an SPT regulatory system. However, this is a fascinating example of how a cell with specialized sphingolipid requirements has developed an ORMDL‐dependent regulation of sphingolipid biosynthesis that serves its needs. Outside the scope of this review is another mechanism by which vascular endothelial cells regulate SPT activity, through the action of Nogo B, a member of the reticulon, membrane shaping family of proteins [62, 63]. Nogo B binds to and inhibits SPT. The interplay between this regulation and ORMDL‐dependent regulation remains, as yet, undetermined.

8. Discussion

Is regulation of SPT activity really all that important? The answer is definitively yes. Let's begin with the ORMDL‐dependent homeostatic regulation of SPT. Deletion of the ORMDLs is embryonic lethal in mice, to start [54]. In human patients, mutations in SPTLC1 or SPTssa impair ORMDL‐dependent regulation of SPT [64, 65]. The SPTLC1 mutations underly a childhood form of amyotrophic lateral sclerosis (ALS) [65] and the SPTssa mutations lead to spastic paraplegia [64]. Why these two mutations, which both have the effect of blunting the ability of SPT to be regulated by cellular sphingolipid levels, have different clinical characteristics is not yet known. As for regulation of the chain‐length of the sphingoid backbone by the alternate subunits SPTLC3 and SPTssb: Deletion of SPTLC3 is protective in a model of cardiac ischemia [66], as cited above. SPTLC3 expression, and the accompanying increase in short chain sphingoid backbones, is deleterious in this case, but no doubt advantageous in others, for example, in the placenta, where SPTLC3 is highly expressed [67]. No functional tests have yet been made of SPTssb deletion. However, this subunit is abundant in skin keratinocytes, and its incorporation into the SPT complex results in high levels of alternate sphingoid base‐containing ceramides comprising the permeability barrier in skin [51]. This is highly suggestive that the SPTssb‐dependent alteration in sphingoid base production is essential for maintaining the physical characteristics of that barrier.

Two ORMDL‐dependent mechanisms regulate levels of SPT activity, direct binding by ceramide to allosterically inhibit SPT and condition‐dependent enhanced depletion of ORMDL from the SPT complex. Why have two? The direct binding of ceramide to the complex is rapid and reversible. One notion is that this rapid regulation reflects the dual nature of ceramide. Ceramide has long been recognized for its role in promoting apoptosis [68]. In many instances, for example, in hepatic steatosis, liver and cardiac disease, elevated ceramide levels are out and out toxic [69, 70, 71]. The toxic nature of ceramide has been capitalized on as an anti‐cancer therapeutic in a recent formulation currently in clinical trials [72]. On the other hand, ceramide is the obligate metabolic intermediate for the formation of sphingomyelin, glycosphingolipids, ceramide‐1‐phosphate, and sphingosine. Notably, deletion of one of the ceramide synthases (Cers2) has severe physiological consequences, including liver and skin pathology and loss of myelin (reviewed in [73]). Clearly, a ceramide pool must be maintained to serve downstream sphingolipid biosynthesis. The direct ORMDL‐dependent regulation of SPT activity by ceramide accomplishes the goal of maintaining a ceramide pool at optimal levels. When the ceramide pool is depleted by the activity of downstream transporters and enzymes, SPT activity will rise to replenish the pool. When the pool threatens to rise to deleterious levels, SPT activity is reduced until ceramide levels are stabilized. In this context, the distinct ceramide responsiveness of the ORMDL isoforms may serve the needs of different cells and tissues with varying sensitivity to ceramide. In cells that possess an exquisite sensitivity to ceramide, ORMDL3, which has the lowest IC₅₀ for ceramide, may predominate. Conversely, in cells with a higher tolerance for ceramide, and potentially an elevated requirement for downstream ceramide metabolites, ORMDL1 or 2 may be included in the complex to maintain an elevated ceramide pool. This may be the case in the myelin‐producing oligodendrocytes in which levels of ORMDL2, the least ceramide sensitive form, rise during the period when the sphingolipids that comprise myelin are being synthesized in quantity [53].

What, then, of the mechanisms that degrade the ORMDLs and eliminate them from the SPT complex? On the surface, it would seem foolhardy to allow SPT activity to run wild, losing its responsiveness to potentially toxic ceramide levels. This may, however, be the point. Under physiological situations where a robust and sustained increase in sphingolipid is essential, ORMDL‐free SPT can ramp up sphingolipid synthesis to maximum levels without the moderating influence of ceramide pools. Such may be the case under experimental conditions in which sustained inhibition of ceramide drastically reduces sphingolipid levels. The cell is desperate to replenish the sphingolipid pool and so turns SPT on at full speed. Similarly, in vascular endothelial cells in which sustained production of sphingosine‐1‐phosphate (S1P) is required, ceramide levels may be of secondary concern. Thus, when S1P levels drop, SPT is turned on full force without ceramide regulation. The physiological conditions that drive accelerated ORMDL degradation have yet to be identified. One expected condition would be during bursts of sphingolipid production, perhaps in terminal differentiation of skin keratinocytes or during the peak of myelin production in oligodendrocytes and Schwann cells, the myelin‐producing cells of the central and peripheral nervous systems, respectively. The accelerated degradation could also serve as a means to switch the ORMDL isoform incorporated into the complex. When cells are required to move from a relatively ceramide‐insensitive state to a ceramide‐sensitive state, for example, they could degrade ORMDL2 and replace it with ORMDL3.

9. Conclusions

As the first and rate‐limiting step in sphingolipid biosynthesis, it is hardly surprising that SPT activity is homeostatically regulated. Less obvious is the discovery that the subunit composition of the complex dictates the chain length, and therefore the physical properties, of the sphingoid backbone. SPT activity is therefore both quantitatively and qualitatively regulated, a unique property among metabolic enzymes.

In some ways the regulation of SPT mirrors that of HMG‐CoA reductase, the initiating and rate‐limiting step in cholesterol biosynthesis. The ultimate products of the pathways initiated by these enzymes are essential but toxic in excess. It is unsurprising, then, that both enzymes are tightly regulated. HMG‐CoA reductase is regulated at every imaginable level: gene expression, translation, degradation, and post‐translational modification (reviewed in [74, 75]. These multiple levels of regulation allow cholesterol synthesis to be responsive to a variety of conditions. The level of SPT activity is certainly regulated by two distinct ORMDL‐dependent mechanisms. This seems to reflect a similar regulatory flexibility. Whether additional levels of SPT regulation remain to be discovered is an open question but would not be unexpected.

It is remarkable that the ceramide‐dependent regulation of SPT is conserved between higher eukaryotes in both the animal and plant kingdoms, and lower eukaryotes. This understanding emphasizes the critical role that sphingolipids play in the life of these organisms and the fundamental importance of regulating their production. The future awaits a fuller picture of the regulation of this beautiful and essential enzyme complex and the impact of that regulation in physiology and disease.

Author Contributions

U.M. and B.W. conceptualized and wrote the paper.

Conflicts of Interest

The authors state that they have no conflicts of interest to declare.

Mahawar U. and Wattenberg B., “Intricate Regulation of Sphingolipid Biosynthesis: An In‐Depth Look Into ORMDL‐Mediated Regulation of Serine Palmitoyltransferase.” BioEssays 47, no. 9 (2025): 47, e70036. 10.1002/bies.70036

Data Availability Statement

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.