Vishwanath R. Lingappa MD, PhD
Synthia H. Mellon PhD
COMPARTMENTATION OF EUKARYOTIC CELLS IN RELATION TO HORMONE SYNTHESIS & RELEASE
A fundamental feature of eukaryotic cells is the presence of a multiplicity of intracellular membrane-delimited compartments. A number of these intracellular compartments are related as components of the secretory pathway (discussed further below). The luminal spaces (ie, those within membrane vesicles) of these compartments are generally oxidizing environments, equivalent to the outside world (see Figure 2-1). Across the membrane from the luminal space of these compartments is the cytoplasm, which is a reducing environment with very different protein content and enzyme activities. All hormones start their biosynthesis in the cytoplasm and must be released into the outside world by processes of secretion in order to exert their biologic effects. As we shall see, compartmentation provides both challenges and opportunities for hormone biosynthesis and regulation.
Figure 2-1. Conceptual framework for thinking about vesicular and nonvesicular modes of hormone secretion. Schematic diagram of a cell, indicating that some hormones are transported in vesicles to the cell surface and released in quanta upon fusion of the vesicle with the plasma membrane, while other hormones are transported directly across the plasma membrane either by specific transporters or by diffusion. Vesicular secretion can be regulated separately at the level of synthesis and of release, while control over synthesis is the only known way to regulate hormones released in a nonvesicular manner. Examples of hormones secreted in a vesicle-mediated fashion are most polypeptide hormones and catecholamines. Examples of hormones released in a nonvesicular manner are the eicosanoids, steroids, and thyroid hormones.
OVERVIEW OF HORMONE BIOSYNTHESIS
Any molecule delivered via the bloodstream that can be recognized by a receptor in or on a target cell in a manner that conveys information can, in principle, be used by the body as a hormone. In the course of evolution, an enormous variety of molecules have been utilized as hormones by various organisms. With this diversity of hormone structure comes a corresponding variety in modes of hormone biosynthesis. In this chapter, we shall summarize current concepts of hormone synthesis and release.
Hormones can be divided into two broad classes (Figure 2-1). The first class consists of those that are stored in membrane vesicles. These hormones are released from an endocrine cell by fusion of membrane vesicles with the plasma membrane, typically in response to a stimulus for secretion that may or may not be coupled with the stimulus for hormone synthesis. Hormones of this class are often synthesized and then stored for later release. After release of stored hormone upon fusion of membrane vesicles with the plasma membrane, the “empty” membrane is recycled by a process known as endocytosis. The second broad class of hormones consists of those secreted immediately upon synthesis in a manner not mediated by membrane vesicle fusion. For these hormones, there is typically no distinction between the stimulus for synthesis and the stimulus for release—hence, control over synthesis is the major means known for regulating their secretion.
The polypeptide hormones comprise the most prominent example of hormones whose release is vesicle-mediated, and their secretion involves transit through multiple membrane-delimited compartments of the classic secretory pathway (see below). However, the vesicle-mediated pathway is also used for the secretion of a number of nonpolypeptide hormones and neurotransmitters such as catecholamines (eg, dopa-mine) and γ-aminobutyric acid (GABA). These small molecules do not undergo the full range of intracellular trafficking events involving transport through multiple intracellular compartments that is seen in protein secretion. These nonpolypeptide hormones are either newly synthesized in the cytoplasm or are taken up by specific transporters from outside the cell and pumped back into recycled empty vesicles in preparation for another round of vesicle fusion and secretion. As will be summarized below, great progress has been made in
recent years in our understanding of the molecular mechanisms of vesicular traffic involved in biosynthesis and release of both peptide and nonpeptide hormones.
Classes of hormones released by non-vesicle-mediated mechanisms include the steroids(glucocorticoids, androgens, estrogens, andmineralocorticoids—all derived from cholesterol) and the eicosanoids (a family of fatty acid-derived signaling molecules that include theprostaglandins). Exactly how non-vesicle-mediated release occurs is not as well understood as vesicular secretion. Historically, simple diffusion was assumed to be the mechanism both for release from the hormone-producing cells and for entry into target cells. In recent years, however, specific transporters have been implicated in directing some of these classes of molecules out of the cell. Whether all non-vesicle fusion-mediated secretion will prove to be driven by transporters through specific channels or whether some hormones leave the cell solely by diffusion remains a subject for future investigation.
MEMBRANE VESICLE-MEDIATED HORMONE EXPORT
With the exception of a very small number of proteins synthesized within the mitochondrial matrix and in plants, within chloroplasts, most proteins made in eukaryotic cells are synthesized on ribosomes in the cytoplasm. Proteins that reside in the cytoplasm are not generally released in the absence of cell damage or death. Thus, specialized mechanisms must exist to discriminate between newly synthesized proteins that are destined to be secreted and those that are to remain in the cytoplasm.
The polypeptide hormones are an important subset of secretory proteins, and the classic secretory pathway by which they leave the cell is the best-understood mechanism of hormone export. It appears to be the mechanism used for most but not all polypeptide hormones and can be separated into early and late events. The early events involve getting the newly synthesized secretory protein into the luminal space of the endoplasmic reticulum (ER), a membrane-delimited compartment. In the course of those early events, the protein must be properly folded and is often covalently modified, eg, by the addition of carbohydrates. The late events involve transport of the properly synthesized, folded, and modified protein from the ER lumen to the lumen of other membrane-delimited compartments, including the Golgi apparatus, and subsequently to secretory granules and ultimately, upon fusion of a secretory granule with the plasma membrane, out of the cell. Whereas the early events involve direct transfer of the individual secretory polypeptides across the lipid bilayer of the ER membrane, in the later events movement of secretory proteins occurs exclusively by transport within a membrane system—or by budding of membrane vesicles from one compartment and their subsequent fusion with another compartment. Thus, the protein cargo moves from vesicle to vesicle without ever again actually crossing a lipid bilayer directly (Figure 2-1).
Targeting to the Membrane of the Endoplasmic Reticulum
The early events of polypeptide hormone targeting to and translocation across the ER membrane generally occur while the protein is still being synthesized. The mechanism that appears to have evolved as the major pathway of ER membrane translocation in eukaryotes involves a sort of molecular “ZIP code” termed the signal sequence, which is found as part of nascent secretory proteins (Figure 2-2). The signal sequence is a sequence of amino acid residues encoded in the mRNA, usually but not always at the 5′ end of the coding region and thus usually occurring at the amino terminal of the encoded protein. Signal sequences typically are composed of three domains, including a stretch of hydrophobic residues. While the hydrophobicity was once thought to relate to interaction with the lipid bilayer, it now seems more likely that it reflects features recognized by other proteins which serve as receptors for signal sequences. Curiously, signal sequences display a substantial heterogeneity from one protein to another. This was once interpreted as meaning that hydrophobicity rather than specific sequence was important for signal sequence function. However, recent findings suggest that even subtle differences in the specific sequence of the hydrophobic domain can have dramatic functional effects due to altered protein-protein interactions (see below).
The signal sequence typically emerges from the ribosome as part of the nascent protein chain and is quickly bound by a cytoplasmic ribonucleoprotein complex composed of six polypeptides and a small RNA, termed signal recognition particle (SRP), that serves to target the nascent chain to the membrane of the ER. “Professional” secretory tissues such as endocrine glands often contain a substantial amount of “rough” ER visible under the electron microscope, so-called because of the presence of ribosomes bound to the ER membrane that have been targeted there through the action of SRP.
In molecular terms, SRP serves in two ways to facilitate translocation of nascent secretory proteins across the ER membrane. First, binding of SRP to both the signal sequence and the ribosome slows the rate of chain elongation, thereby increasing the“window of
time” during which the chain can find the cytoplasmic face of the ER while still nascent and thus still able to be translocated into the lumen. Most proteins cannot be translocated after their synthesis has been completed and they have been released from the ribosome. In part this may be because most nascent chains need to be at least partly unfolded for translocation. The cotranslational requirement may also in part reflect a role for the ribosome in opening the channels through which translocation occurs (see below).
Figure 2-2. Translocation of nascent polypeptide hormones across the ER membrane. Key steps in polypeptide hormone biosynthesis and secretion are indicated by the circled numbers. (1) Ribosomes from the free cytoplasmic pool initiate translation of mRNA encoding secretory proteins. A protein complex termed nascent chain associated complex (NAC) appears to prevent inappropriate or premature targeting of the nascent chain. (2) The initial codons encode a signal sequence that emerges from the ribosome and will serve to target the nascent secretory chain to the ER membrane by virtue of the affinity of a cytosolic particle termed signal recognition particle (SRP) for both the signal sequence and a receptor in the ER membrane. SRP binding displaces NAC from the nascent chain; SRP docking to the ER membrane displaces the nascent chain from SRP. (3) Assembly of a channel across the ER membrane provides a pathway for the nascent growing chain to enter the ER lumen. Translocation involves a ribosome-membrane junction sufficiently “tight” to shield the nascent growing chain from the cytoplasm, thereby preventing the cytoplasm from being secreted into the ER lumen. (4) During translocation, the chain is subject to a host of cotranslational modifications including the addition of carbohydrates which are transferred as a preformed structure composed of 11 carbohydrate residues from lipid carriers to the protein. (5) Translocation-associated events in polypeptide hormone biogenesis usually include cleavage of the signal sequence. (6) As the chain emerges into the ER lumen, folding occurs, governed by molecular chaperones in the ER lumen. (7) After completion of protein synthesis, the ribosome subunits are released back into the free cytosolic pool, the channel disassembles or closes, and the folded chain is localized to the ER lumen.
Second, when bound to a signal sequence, SRP undergoes conformational changes resulting in high affinity for a receptor on the ER membrane (termed the SRP receptor), thereby targeting the nascent chain to the correct subcellular location. It would be inefficient, for example, to target a secretory protein to the mitochondria by accident.
Upon binding the SRP receptor, SRP undergoes another conformational change that releases the signal sequence and the ribosome, allowing the rate of chain
elongation to increase and freeing the signal sequence and nascent chain to interact with other proteins in the ER membrane. During the same time, a functional ribosome membrane junction forms by interaction with receptor proteins.
Some of the proteins that comprise both SRP and the SRP receptor are GTP binding proteins, and it appears that cycles of GDP-GTP exchange and GTP hydrolysis serve to ensure the fidelity of targeting and the unidirectionality of several of the early events in translocation, including at least (1) signal recognition by SRP, (2) targeting of the nascent ribosome-signal sequence-SRP complex to the SRP receptor, (3) assembly of the translocation channel, and (4) release of SRP to participate in another round of targeting.
Finally, a non-SRP-mediated pathway of targeting has been described in yeast. Thus, SRP-deficient yeast mutants are viable but grow slowly. Some proteins in these mutants are severely affected by lack of SRP while others appear largely unaffected, presumably because they can target to the ER membrane as efficiently by non-SRP-mediated mechanisms. The full significance of this finding remains to be established. However, it is consistent with the concept that crucial biologic recognition systems have evolved a degree of redundancy in case (for whatever reason) the primary recognition system should fail.
Translocation Across the ER Membrane
Once targeted, the nascent chain must somehow traverse the ER membrane, which is otherwise a barrier to movement of most proteins from the cytosol to the ER lumen. Upon release from SRP, the correctly targeted signal sequence appears to interact with other proteins in the ER membrane to open a channel to the ER lumen. The growing polypeptide chain—but not other proteins or even the ions in the cytosol—has access to the channel. The nascent chain is translocated through this channel into the lumen of the ER, tightly shielded from the cytosol by the ribosome-membrane junction. It appears that there are “gates” at each end of the translocation channel. Recent studies suggest that another form of regulation involves transient opening of the ribosome-membrane junction directed by particular sequences present in some nascent proteins. As a consequence, translocation across the ER membrane is interrupted and specific domains of nascent secretory proteins are exposed to the cytoplasm prior to reestablishment of a tight ribosome-membrane junction and resumption of translocation (see below).
In addition to the GTP hydrolysis-dependent steps of targeting, subsequent steps of translocation also appear to involve ATP hydrolysis. The precise role of the energy requirement of chain translocation—apart from that involving GTP hydrolysis for targeting—remains unclear. Perhaps the energy of ATP hydrolysis is needed to form or maintain the channel in an open conformation, or to actually pull the chain across the membrane, as has been suggested for protein secretion in bacteria. Alternatively, ATP may be hydrolyzed as part of the action of molecular chaperones acting on the chain in the ER lumen, and the actual translocation may even occur by brownian motion without any requirement for ATP hydrolysis.
Major progress has been made in recent years in identifying ER membrane proteins involved in nascent chain translocation across the ER membrane. It appears that at least two membrane protein complexes, the heterodimeric SRP receptor and the heterotrimeric Sec 61p complex, are involved in translocation of essentially every secretory protein. A third membrane protein, termed translocating chain-associated membrane protein (TRAM), is involved in translocation of most but not all secretory proteins. Roles for other accessory proteins in translocation of specific subsets of polypeptides or in more complex modification events associated with translocation of particular proteins have been demonstrated in some cases. Together, the complex of proteins involved in both translocation and modification of newly synthesized secretory proteins is termed the translocon. Recent studies have strongly suggested that the translocon in eukaryotes is an aqueous protein-lined, protein-conducting pore. Yet lipids appear to have at least transient access to the nascent chain during its translocation, perhaps reflecting the dynamic and transient nature of a channel rapidly assembled and disassembled from component subunits. Upon completion of secretory protein synthesis, the ribosomal subunits release from the cytoplasmic side of the ER membrane and the transmembrane channel disassembles or in some other way closes, leaving the newly synthesized secretory protein localized to the luminal space of the ER with no way to return to the cytosol (Figure 2-2).
Co- & Posttranslational Modification of Proteins
During translocation across the ER, subsequently within the ER lumen, and in a variety of later membrane-delimited compartments, newly synthesized proteins may be subject to literally dozens of possible processing events that often vary from protein to protein, tissue to tissue, and species to species. Some of these modifications are covalent—eg, proteolytic removal of the signal sequence, disulfide bond formation, and addition of various moieties to amino acid side chains (eg, carbohydrates). Other modifications are noncovalent, such as proper folding of newly synthesized polypeptides
under the supervision of families of proteins termed molecular chaperones that are present in the cytosol, ER membrane, and ER lumen. Only a small number of the many possible modifications that are possible actually occur on any given specific protein, and some proteins appear to receive no modifications at all. Furthermore, even for those specific proteins which receive covalent modifications, the modifications are heterogeneous. That is, some copies of the protein receive the full complement of modifications, while other copies receive only some or even no modifications at all. A body of work suggests that these differences are not random or stochastic events but rather may be regulated and of functional significance for hormone action (see below). For the few modifications that have been well studied, it appears that either primary amino acid sequence motifs or secondary or tertiary structural features of a particular newly synthesized protein are recognized by the enzymes that carry out these modifications.
Perhaps the best-understood posttranslational modification is a subset of carbohydrate addition termed N-linked glycosylation (Figure 2-2). This form of glycosylation occurs on selected asparagine residues during translocation of the nascent chain into the ER lumen. The carbohydrate units consist of up to 14 sugars that are assembled in a tree-like structure on lipid transporters called dolichols. The entire sugar tree is then transferred en bloc to selected asparagine residues of the nascent protein as it enters the ER lumen (hence the term “N-linked” glycosylation). Subsequently, the individual sugars that comprise the sugar tree are modified, with some removed and others added, in the ER and in more distal compartments of the secretory pathway.
In most cases, the precise functions of the various covalent co- and posttranslational modifications, including N-linked glycosylation, are unknown. However, as will be discussed below, two important roles for N-linked carbohydrates in facilitating proper sorting and traffic of some proteins through the secretory pathway have been discovered. In other cases, changes in carbohydrates have been shown to alter the activity of particular hormones once secreted, either by affecting the affinity of binding to hormone receptors or by altering the clearance from the bloodstream and hence the half-life and effective concentration of the particular hormone in blood. An important frontier of hormone biosynthesis and action is to understand the significance of the observed heterogeneity in modification of hormones and how it is regulated.
Quality Control by the ER
Within the ER lumen, molecular chaperones not only facilitate proper folding of the newly arrived polypeptide, they also assess the outcome of folding. Often, as a result of interaction with molecular chaperones, proteins deemed improperly folded are not allowed to leave the ER even though, in at least some cases, those proteins appear sufficiently well folded to be capable of physiologic function in vitro. This “quality control” function appears to be a major role of the ER in the scheme of protein biogenesis. Whether the molecular chaperones that carry out these recognition events for protein folding and quality control also play a role in the actual translocation of proteins into the ER lumen remains controversial. For example, the protein BIP (immunoglobulin-binding protein) has been implicated by some work as a “plug” preventing proteins from leaking back into the cytosol. Other work suggested that BIP may be part of a “molecular ratchet” pulling proteins into the ER lumen via the translocation channel.
In the case of one particular ER molecular chaperone, termed calnexin, trimming of the glucose units at the end of the N-linked carbohydrate tree appears to serve as a monitor of the need for further chaperone action for at least some newly synthesized proteins. Calnexin binds these unfolded chains upon trimming of the outer two of the three glucose residues found on the N-linked carbohydrate tree. Like other molecular chaperones, calnexin uses the energy from cycles of ATP hydrolysis to prevent misfolding of the newly synthesized protein. Upon completion of proper folding, the final glucose is trimmed, calnexin is released, and the protein is transported in vesicles out of the ER. If proper folding, as defined by the quality control machinery, is not achieved, glucose may be added back and multiple cycles of calnexin binding, folding, sugar trimming, and release allowed to occur. Calnexin is unusual among the molecular chaperones in that it recognizes the polar carbohydrate moiety. In the case of most other molecular chaperones, hydrophobic interactions seem to be the important determinants of interaction with substrates.
In a number of cases, proteins deemed to be improperly folded by the quality control machinery are not only prevented from leaving the ER, they are rapidly degraded as well. Some data suggest that in the case of some particularly complex secretory and integral membrane polypeptides, rapid degradation in the ER can be a point of regulation of protein biogenesis, multisubunit assembly, and secretion. At least one of the mechanisms of rapid degradation of newly synthesized proteins in the ER involves reverse translocation through the Sec 61p complex, conjugation to the cytoplasmic protein ubiquitin, and subsequent degradation in the cytoplasm by the proteasome.
What keeps the different compartments of the cell “on the same page” with each other in terms of maintaining the correct levels of activity for synthesis, maturation, degradation, etc? The unfolded protein response
is a feedback loop, highly conserved from yeast to mammals, that serves as a model for such coordination. Accumulation of unfolded proteins in the ER results in activation of a kinase that signals to the nucleus the need to up-regulate transcription of genes encoding molecular chaperones involved in protein folding, thereby correcting the deficiency of molecular chaperones that resulted in unfolded protein accumulation in the first place. Recent work has implicated this pathway in the degradation of proteins as well. It is likely that many aspects of intracellular coordination between organelles are governed by similar feedback loops yet to be discovered.
The aforementioned summary of protein biogenesis, from synthesis and translocation to degradation or export out of the ER, represents the currently accepted general framework for the early events of the secretory pathway. However, many details relevant to endocrinology may not be fully explained by the simplest version of this paradigm. In particular, the heterogeneity of protein modifications on specific proteins is intriguing. Why does this heterogeneity exist if it is not significant? If significant, how is it achieved and regulated?
Many hormones have been implicated in multiple functions, but often only one major function has been well studied. Might heterogeneity of modification or folding of newly synthesized polypeptide hormones provide a basis for generating subsets of hormone molecules that are responsible for nonclassic hormone actions? Very recently there has been some insight into the regulation of pathways of maturation, which suggests that signal sequences previously thought to be involved only in targeting and translocation, as described above, may also play a role in selecting the precise pathway of maturation achieved by a given nascent chain. The mechanism by which signal sequences do this appears to be altering of the structural organization of the translocation channel through which the nascent protein traverses the ER membrane. This has been demonstrated for signal sequence mutations—but as yet unidentified protein-protein interactions involving the signal sequence could presumably have a similar effect. Thus, two copies of the identical protein traversing translocation channels of different organization could achieve different modifications or different folding and therefore could perform different functions. The precise mechanism by which this process is regulated remains to be fully understood; however, some data suggest that signal peptidase, the multicomponent enzyme which removes the signal sequence from the nascent chain by endoproteolytic cleavage, may be involved.
These findings have led to a provocative new hypothesis which suggests that many chains conventionally viewed as “misfolded”based on their rapid degradation in the ER under normal circumstances may in fact represent alternately folded forms with distinctive functions not wanted by the cell except at particular times—which are extremely transient and therefore hard to detect. Degradation in these cases is therefore not due to misfolding but occurs because the alternately folded forms are not desired at that particular time. Whether this phenomenon is general or is utilized by a few specialized proteins remains to be determined, though some recent data support this hypothesis when applied more generally to polypeptide hormones and their receptors. A protein that can be folded in two different ways both of which are functional is tantamount to being two different proteins—ie, each folded form has a different shape as seen by receptors in the “outside world”. Thus, regulated folding of secretory and membrane proteins would have enormous implications for hormones and for the information content of the genome if it proves to be a phenomenon of many proteins.
Post-ER Vesicular Traffic in the Secretory Pathway
Translocation across the ER membrane is the only time that a secretory protein directly crosses a lipid bilayer. All subsequent trafficking steps involve movement within the confines of a membrane compartment (vesicular traffic), ie, the pinching-off of a membrane vesicle containing a cargo of newly synthesized protein from one membrane-delimited compartment and its transfer to another membrane compartment by fusion of the vesicle. Ultimately, vesicles fuse to the plasma membrane (exocytosis) or to the lysosome (lysosomal targeting; Figure 2-3). Concomitantly with traffic of vesicles through the secretory pathway, a recycling of membrane must occur. Both vesicular and tubular pathways of membrane recycling have been proposed (Figure 2-3).
The exact number of membrane subcompartments in the secretory pathway has yet to be determined. Indeed, in some cases the distinction between one compartment and the next may be largely semantic. At the least, all secretory proteins are transported in vesicles from their site of synthesis in the ER to a post-ER “intermediate” compartment from which vesicles deliver them to the Golgi apparatus. The intermediate compartment serves as a kind of recycling center from which proteins that should remain in the ER or which are not competent to progress down the secretory pathway are returned to the ER. Within the Golgi, they are transferred serially from the so-called cis-Golgi network to the medial Golgi stack, to the trans-Golgi stack, and finally to the trans-Golgi network (TGN) membrane cisternae. Each serial compartment is operationally defined
by localization of specific enzymes (such as those involved in modifying carbohydrates) to a particular subset of membranes. As more compartment-specific genes are cloned and monospecific antibodies generated to modifying enzymes, it is likely that additional subcompartments will be operationally distinguished within the currently understood secretory pathway. However, not every protein which traverses the secretory pathway receives a modification in any particular compartment as far as is known.
Figure 2-3. Membrane traffic in hormone-secreting cells. Polypeptide hormones travel through a variety of membrane compartments of the secretory pathway, with transport mediated by the fission and fusion of vesicles. Newly synthesized polypeptide hormones exit the ER and travel in vesicles to the intermediate compartment where they are concentrated by a number of different mechanisms, including recycling of “empty” containers back to the ER. Vesicles transport the proteins through the Golgi apparatus to the trans Golgi network whence other vesicles transport the proteins to either the regulated or constitutive secretory pathways or to lysosomes for degradation. In addition to forward transport, a pathway of retrograde transport, perhaps largely mediated by tubules rather than vesicles, returns membrane and some proteins to earlier compartments, including the ER and Golgi stacks. Endocytosis is the set of membrane-delimited trafficking pathways whereby hormones bound to receptors at the cell surface are internalized and transported either to the lysosome for degradation or back to the cell surface for re-release. The pathways of endocytosis and biogenesis can overlap in key compartments and share common features of mechanism—eg, the role of “coat” proteins in forming and targeting vesicles from trans Golgi network to plasma membrane during protein biogenesis and from plasma membrane to lysosome in endocytosis.
Trafficking of proteins through the vesicular stages of the secretory pathway involves recognition events both in the lumen, such as that described above for calnexin, and in the cytosol. The cytosolic recognition events involve cytoplasmically disposed domains of proteins within the vesicular membrane. Various families of small and large GTP-binding proteins have been implicated in control of these cytosolic recognition processes.
In a manner analogous to the mechanism targeting the nascent secretory protein to the ER membrane described above, cycles of GTP for GDP exchange followed by GTP hydrolysis ensure correct vesicle formation, docking, and fusion to the correct target compartment. Other GTP-binding proteins confer unidirectionality on the membrane fusion event, which is itself mediated by a“fusion machine” assembled from various proteins including some in the cytosol and some that were also involved in vesicle formation and targeting.
Figure 2-4. SNARE hypothesis for intracellular vesicular traffic. Vesicular traffic requires a means of discrimination on both sides of a membrane so that (1) products destined for one fate can be sorted from those with another, and (2) so that the sorted molecules can be correctly targeted to the correct subsequent membrane compartment. This includes a role for coat proteins that associate with the surface of a compartment to form or pinch off vesicles. (3) Once vesicles have formed—by budding from the so-called donor membrane—their trafficking is governed by recognition events between molecules termed v-SNAREs on their surface with molecules termed t-SNAREs on the surface of the target membrane. Before localization of cargo to the target compartment can be consummated, several events must occur in concert with v-SNARE-t-SNARE interaction, including uncoating (4), docking (5), and fusion (6). These events can be constitutive or regulated simply by placing one or more of them under control of a signal-transduction pathway.
Differences in adapter proteins (including some that bind and hydrolyze GTP) and surface receptor proteins termed v-SNARES with cognate binding sites (termed t-SNARES) on the correct membrane of destination are believed to mediate the correct targeting of vesicles throughout the secretory pathway (Figure 2-4).
In some cases, various “coats” composed of specific proteins form on the outside of membranes, allowing them to be pinched off to form vesicles. A protein called COP II is found on vesicles that bud from the ER destined for the intermediate compartment. A protein called COPI is found on vesicles that return from the intermediate compartment back to the ER. Vesicles of different sizes, containing different classes of proteins as their cargo and destined for different subcellular fates, are often observed to have different coats or different adapter proteins that mediate assembly of a particular coat, permitting the resulting vesicle to be selectively targeted to the correct membrane destination. The TGN is the last compartment of the Golgi apparatus, from which vesicles are targeted to a variety of locations, including the regulated and constitutive secretory pathways and the lysosome (see below). Phosphatidylnositol transfer proteins greatly stimulate the formation of vesicles from the TGN. These negatively charged phospholipids may play a role in vesicle formation, perhaps
by inducing membrane curvature or by recruiting coat proteins and other cytosolic components.
Membrane Trafficking Steps Prior to Sorting of Cargo
Through the TGN, the cargo contained within the vesicles that move from compartment to compartment is nonspecific, meaning that no sorting of content proteins has occurred. Thus, for example, a vesicle leaving the cis Golgi and targeted to the medial Golgi will contain a mixture of newly synthesized proteins including enzymes to be localized within the lysosome (an acidic membrane-bound organelle containing various hydrolytic enzymes that are active at low pH), proteins to be secreted continuously, and proteins to be secreted only in response to specific stimuli (ie, hormones). Likewise, the membrane of this vesicle contains various classes of membrane proteins with different destinations (eg, those targeted selectively to the apical plasma membrane and others targeted exclusively to the basolateral plasma membrane). While the cargo delivered to the TGN is nonspecific, the pathway followed by the vesicles is not. A vesicle derived from the cis Golgi must be selectively targeted to the medial Golgi and not the trans stack or back to the ER or to the plasma membrane. A separate pathway of vesicular transport is involved in “recycling” of membrane containers by return from cis Golgi to ER, etc. This recycling pathway is also used to return so-called resident proteins (such as BIP and calnexin) that need to be retained in a particular compartment such as the ER. It appears that at least one of the mechanisms involved is not retention per se but rather efficient retrieval of wayward polypeptides from the subsequent compartment. The coats of the intra-Golgi vesicles with nonspecific cargo and involved in recycling also contain a protein COP I. The coats of vesicles involved in ER to intermediate compartment and cis Golgi traffic contain the protein COP II.
Despite the nonspecific nature of their cargo, vesicle recognition events are probably occurring within the lumen of the ER and Golgi stack. For example, it was recently noted that some membrane proteins are concentrated in the intermediate compartment between ER and Golgi, strongly suggesting the occurrence of selective luminal recognition events. Whether this is true for all proteins or only for a subset within the nonspecific cargo remains to be determined.
A large body of evidence, summarized above, supports the notion that transfer from cis to medial to trans Golgi and TGN cisternae involves budding of vesicles from one compartment and their fusion to the next compartment within the Golgi apparatus. However, controversy has emerged as to the relative role of vesicles versus tubules of membranes that may connect one compartment with the next and whether one versus the other mechanism is involved in “forward” transport versus the return of “empty” membrane containers back to their compartment of origin. Recent work suggests that the concept of relatively fixed compartments of the Golgi complex, between which newly synthesized proteins are moved by vesicular traffic, may have to be at least partially revised in favor of a more dynamic “maturational” model in which cisternae progress from cis to medial to trans.
Sorting of Vesicle Cargo at the TGN
Unlike most vesicular transport from ER to TGN, true sorting occurs upon exit from the TGN. Vesicles leaving the TGN are heterogeneous with respect to both their cargo and their destination. Some vesicles are loaded with exclusively lysosomal enzymes and are targeted selectively to an acidic endosomal compartment in a pathway that will lead eventually to the lysosome or to a compartment which becomes a lysosome.
Other vesicles lead to the regulated secretory pathway, where the final vesicular fusion event of a secretory granule fusing to the plasma membrane to release the secretory proteins happens only upon specific stimulation. Thus, the stimulus to insulin release via the regulated secretory pathway is hypoglycemia; that of parathyroid hormone release is hypocalcemia; that of renin release includes adrenergic neural stimulation; etc. Some of the vesicles that are released by the regulated secretory pathway first condense their contents into a concentrated precipitate of secretory product and proteoglycans. These vesicles are termed secretory granules (see below). Other regulated secretory vesicles are morphologically distinguishable in that their contents do not undergo concentration. However, they do display stimulus-secretion coupling, the sine qua non of the regulated secretory pathway.
Still other vesicles, including those containing many membrane proteins, enter the constitutive secretory pathway. The constitutive pathway differs from the regulated pathway in that nothing prevents the final vesicle from fusing with the plasma membrane as soon as the vesicle has formed and been targeted—ie, no stimulus is required to overcome a preexisting block that prevents vesicle fusion to the plasma membrane in the absence of stimulus.
In addition to segregating content proteins with different subsequent vesicular destinations, the TGN sorting step sets in motion another set of posttranslational modifications. For example, the proteolytic processing of proinsulin to insulin and of pro-opiomelanocortin (POMC) into ACTH and other active peptides is initiated in the TGN and continues in the vesicles targeted to the regulated secretory pathway (see below).
The sorting event that occurs at the TGN is probably mediated by similar interactions on the cytosolic side of the vesicles, as just described for the vesicular transport of nonselective cargo—as well as by additional interactions that specifically connect the sorting event occurring on the luminal side with a particular targeting event on the cytosolic side. The best-defined of these TGN sorting events is that involved in diverting lysosomal enzymes from the common secretory pathway to the lysosome.
Lysosomal Enzyme Sorting
The lysosome is a membrane-delimited organelle in which various hydrolytic enzymes with an acid pH optimum are found. These enzymes—proteases, DNases, RNases, lipases, and the like—are used to break down macromolecules into recyclable building blocks that can be transported to the cytosol or elsewhere for reuse.
The lysosomal hydrolases that ultimately reside within the lysosome start out in the ER lumen, where they are glycosylated along with many other secretory proteins. Once in the cis Golgi, however, lysosomal enzymes are specifically recognized by an N-acetylglucosamine (GlcNAc) phosphotransferase, resulting in the addition of a GlcNAc phosphate moiety to the terminal mannose residue of their N-linked carbohydrate tree (exposed upon trimming of the glucose residues discussed above). Unlike the signal sequence, the recognition of lysosomal enzymes by GlcNAc phosphotransferase is not based on binding of a linear sequence of amino acid residues but is rather a function of affinity for a so-called signal patch composed of amino acid residues from different parts of the molecule that come together upon three-dimensional folding of the protein. This is an important experimental point since a single linear peptide is easy to define and manipulate, while a targeting-sorting signal composed of disparate parts of a protein is more difficult to analyze.
A subsequent enzyme removes the GlcNAc, exposing mannose 6-phosphate, which specifically binds a luminally disposed domain of the transmembrane mannose 6-phosphate receptor found in the TGN. The cytosolically disposed domain of the transmembrane mannose 6-phosphate receptor binds a specific adapter protein that catalyzes coat formation, vesicle budding, and targeting toward a prelysosomal, early endosomal compartment. Upon fusion with this membrane, the mannose 6-phosphate receptor and the lysosomal enzymes it has bound are exposed to a slightly acidic environment in which mannose 6-phosphate—and therefore the lysosomal enzymes—has much lower affinity for the M6P receptor, resulting in dissociation of lysosomal enzymes from the mannose 6-phosphate receptor. Once dissociation has occurred, the receptor is free to recycle back to the TGN, leaving the lysosomal enzymes in the endosomal compartment, which can either fuse to late endosomes and ultimately to lysosomes or “mature” into first a late endosome and then into a lysosome. Recent data support the maturational model for the relationship of endosomes to lysosomes.
Note that the presence of an intermediate acidic compartment between TGN and lysosomes allowed recycling of mannose 6-phosphate receptors without subjecting them to the potentially damaging lysosomal environment. While the mannose 6-phosphate receptor pathway provides a framework for understanding one possible way in which sorting occurs, it is likely not to be the only way to sort lysosomal enzymes since patients with I cell disease, in which a mutant GlcNAc phosphotransferase fails to tag lysosomal enzymes, are observed to secrete their lysosomal enzymes from only some tissues (eg, fibroblasts but not liver). In liver and other tissues, lysosomal enzymes are correctly localized to the lysosome despite lack of the M6P tag, suggesting the existence of an alternative pathway for lysosomal enzyme recognition and sorting.
Much of the recent progress in understanding the pathway of regulated secretion by which most hormones are released on specific stimulation comes from work on the molecular events of synaptic transmission, which can be viewed as a specific example of regulated secretion. Synaptic transmission occurs very quickly, placing limits on the number and nature of interactions, including simple diffusion, that can occur between stimulus and secretion. A number of proteins in the synaptic vesicular membrane have been identified and cloned, and functions have been ascribed to most. Many of these proteins are also found in neuroendocrine cells and are not exclusively neuronal proteins.
In both neurons and neuroendocrine cells, regulated secretion is dependent on an elevation of intracellular calcium. However, calcium probably acts at a number of different steps in each, involving both high- and low-affinity calcium receptor proteins. These steps probably include priming or docking of vesicles as well as vesicle fusion itself. They may also include events such as local disassembly of cytoskeletal proteins to allow secretory granules to dock at the plasma membrane. To date, the precise differences between neuronal and neuroendocrine-regulated secretion have not been fully resolved. More importantly, however, it is now recognized that a number of proteins are involved in common in both of these two modes of regulated protein secretion.
Some of the proteins involved probably play a regulatory role but appear not to be required for exocytosis
per se, based on studies of transgenic knockout mice and drosophila and Caenorhabditis elegans mutations. This group includes the proteins Rab3a, unc-18, synapsin, and synaptotagmin. The latter protein family is believed to comprise the calcium sensors that trigger exocytosis. These conclusions, however, are clouded to some extent by the possibility that multiple genes encoding functionally related products exist and that another gene product is at least partially able to offset the lack of a deleted or mutated gene.
Other proteins are clearly essential components of the universal machinery for vesicular exocytosis. These include gene products specifically cleaved by the clostridial neurotoxins that inhibit regulated secretion from both neurons and neuroendocrine cells. Synaptobrevin, syntaxin, and synaptosome-associated protein 25 are examples of this second group.
Finally, a number of proteins such as N-ethyl maleimide-sensitive factor (NSF) and SNAPs (soluble NSF-attachment proteins) clearly associate with the neurotoxin substrates. Based on this association, synaptobrevin, syntaxin, and synaptosome-associated protein 25 can be viewed as SNAP receptors (SNAREs). Thus, regulated secretion in both neurons (synaptic transmission) and in neuroendocrine cells can be unified under the SNARE hypothesis described earlier for other steps of vesicular transport (Figure 2-4). The novel twist in this hypothesis to accommodate regulated secretion is that the ability of the secretory vesicle v-SNARE to dock with the plasma membrane t-SNARE is dependent on the calcium-signaling step that triggers secretion. In the next few years, this hypothesis is likely to be tested, and if it is fully validated in either neuronal or neuroendocrine systems, a number of unanswered questions can be resolved.
The above general description of regulated secretion is modified to one extent or another in different systems, most likely representing fine-tuned adaptations for homeostasis. For example, in the hypothalamic-pituitary-end organ axis, regulated secretion from the hypothalamus is “pulsatile,” with a distinctive amplitude and rate of release of the secretory products. Perturbation of these parameters in the hypothalamus can affect function at every step of the axis. As another example, in the B cells of the pancreatic islets of Langerhans, insulin secretion has a rapid phase and a slower, more delayed phase—both occurring sequentially in response to the secretory stimulus of an elevated blood glucose. The initial rapid phase is believed to exhaust a readily secretable supply of insulin, which is subsequently replenished, giving rise to the slower second phase. From recent work, it appears that glutamate released from mitochondria in the B cell plays an important intracellular signaling role in coordinating the reloading of the “launch bays” for the second phase. Likewise, ATP-sensitive potassium (KATP) channels are inhibited by intracellular ATP and activated by ADP control exocytosis of insulin in pancreatic B cells. Nutrient oxidation in B cells leads to a rise in the [ATP]-to-[ADP] ratio, which in turn leads to reduced KATP channel activity and plasma membrane depolarization. This in turn activates voltage-dependent Ca2+ channels, which triggers Ca2+ entry and exocytosis.
Regulation of Hormone Release After Exocytosis
Upon condensation into secretory granules, the vesicle content exists as a crystalline precipitate and is no longer osmotically active, a property that facilitates their storage in the regulated secretory pathway. Upon fusion of such a secretory granule to the plasma membrane, the insoluble content must dissolve into the extracellular medium. It appears that the ionic composition of the extracellular fluid can significantly influence the rate of phase transition of the insoluble proteoglycan matrix and hence dissolution and effective release of secretory products. This phenomenon is likely to be an important point of regulation in systems where pulse frequency and amplitude of secretion are important properties (eg, release of neurotransmitters).
POMC Processing & Secretion
There has been considerable progress in understanding the nature of the sorting signal on an individual hormone that accounts for how those molecules are segregated in the TGN from lysosomal, constitutive secretory, and other pathways. In the case of POMC, a specific conformational motif responsible for sorting to the regulated secretory pathway has been identified. It is composed of a 13-amino-acid amphipathic loop close to the amino terminal of the polypeptide and appears to be stabilized by a disulfide bridge. This feature is presumably recognized in the lumen by transmembrane receptors whose cytoplasmic tails recognize adaptin-type proteins responsible for coat assembly, vesicle formation, and proper targeting to the appropriate domain of the plasma membrane.
During this transport process, the vesicles “mature” into secretory granules. Key features of this maturation step include proteolysis (eg, in the case of POMC, where a small peptide, ACTH, will be released from the much larger precursor) and concentration. The concentration step is such that often the regulated secretory cargo forms crystalline arrays of precipitated protein within the granule lumen. Upon stimulation, when the granule fuses to the plasma membrane, the exocytosed product is diluted and dissolves into the bloodstream.
Endocytosis & Recycling
In addition to transport and correct localization of newly synthesized hormones and receptors, vesicular membrane trafficking events are involved in important responses to hormones. In parallel with the pathway of membrane vesicles from ER to plasma membrane is a pathway that leads from the plasma membrane back to various compartments, including the TGN and perhaps even the ER. In part this represents the need to recycle membrane. Note that the volume of regulated secretion in endocrine cells is such that the entire intracellular membrane system would be consumed and localized to the plasma membrane literally in minutes if there were no way to recycle the “empty” membrane containers after use. Not only is there such a pathway, but it probably operates in reverse between each forward compartment (ie, from plasma membrane to TGN, from TGN to trans Golgi, from trans to medial Golgi, from medial to cis Golgi, from cis Golgi to intermediate compartment, and from intermediate compartment to ER).
However, endocytosis is not simply a means of recovering “empty” containers. It provides a means for the cell to take up valuable nutrients and to “sample” the environment. Such is the case for endocytosis of low-density lipoproteins via LDL receptors that are internalized as coated vesicles from coated pits on the plasma membrane of hepatocytes.
Finally, of particular interest for endocrinology, endocytosis provides a means of responding to hormones and other signals. Endocytosis of a vesicle from the plasma membrane places hormone receptors in a protected space where they can no longer be activated by external hormone. This is one (of several) molecular bases of receptor down-regulation and tachyphylaxis to drugs.
Whereas some receptors enter the coated pits only when bound to ligand (eg, EGF receptor), others enter constantly whether ligand-bound or not. In some cases, both ligand and receptor are targeted from the endosome to the lysosomes and degraded (eg, EGF and its receptor). In other cases, dissociation of ligand from receptor in an intermediate acidic compartment allows the receptor to be recycled and only the cargo to be sent to the lysosome (eg, LDL receptor).
One difference from the biosynthetic flow of membranes is that different machinery is involved in vesicle formation from the plasma membrane, which is much more rigid than intracellular membrane compartments by virtue of having cytoskeletal elements bound on the inside and extracellular matrix bound on the outside. Thus, the microtubule-associated protein dynamin has recently been implicated in driving vesicular invagination, constriction, and pinching off from the plasma membrane.
Recently, noncoated invaginations termed caveolae have been noted to occur at the plasma membrane and have been implicated in such hormone-related events as translocation of the insulin-sensitive isoform of the glucose transporter to the cell surface upon insulin stimulation. As mentioned earlier, a complicating issue in vesicular dynamics is whether different gene products can at least partially perform each other's function, thus obscuring the consequences of knockout experiments to assess the role of a particular protein in trafficking events. A better understanding of these features of membrane vesicular dynamics is likely to emerge in the coming years. Some of the differences in pathways defined by the presence or absence of a particular protein may prove to be largely semantic owing to complementary functional activity of another protein despite morphologic, structural, or antigenic differences. In other cases, differences in protein composition may reflect important mechanistic differences in the pathways of vesicular traffic. The importance of recent progress is that such differences in trafficking pathways are no longer the subject of speculation but rather are being detected and studied and the genes associated with them cloned and their expression manipulated.
Secretion of Catecholamines, GABA, & Other Nonpolypeptide Neurotransmitters
Nonpolypeptide products that are released by fusion of vesicles such as catecholamines, acetylcholine, and GABA represent a variation on the theme of recycling through the endocytotic pathway. The vesicles containing these neurotransmitters are loaded with their product by transporters that engage in direct uptake from the cytosol. Loaded vesicles fuse with the plasma membrane in a regulated fashion coupled with a stimulus and usually mediated by a rise in intracellular free calcium. Subsequently, the membrane container is reendocytosed, reloaded, and subjected to another cycle of triggered fusion and content release, as with other forms of stimulus-coupled secretion. Several members of a related family of transporters involved in uptake of these small molecules from the cytosol have been identified.
Secretion of Thyroid Hormone
Thyroid hormone is an example of a hormone whose biogenesis involves a hybrid of vesicular and nonvesicular trafficking. It is synthesized initially as a large polypeptide precursor, termed thyroglobulin, which is assembled into a dimer, iodinated, and modified in the classic secretory pathway before constitutive exocytosis across the apical plasma membrane. Iodinated thyroglobulin
is stored outside of the cell in the thyroid follicle lumen and is termed colloid. Upon appropriate stimulation by thyroid-stimulating hormone (TSH), colloid is taken up by endocytosis and transported across the cell in vesicles that fuse to the basolateral plasma membrane. This variant of endocytosis, in which a product is taken up and delivered from one side of a cell to the other in a polarized fashion, is termedtranscytosis. During transcytosis, the thyroglobulin is degraded by proteases to release active thyroid hormone. Remarkably, out of approximately 6600 amino acid residues comprising the thyroglobulin dimer, six pairs of iodinated, modified tyrosines are released as thyroid hormone, containing either three or four iodines (T3 or T4). Thyroid hormone released from the large polypeptide precursor during transcytosis is able to cross the vesicular membrane and the plasma membrane directly and hence is released immediately upon synthesis in a non-vesicle-mediated manner.
Relevance of Membrane Traffic to Disease
Given the complexities of hormone biogenesis and the many gene products involved in synthesis, maturation, and trafficking of hormones, it should not be surprising that defects in membrane trafficking are prominent among those genetic diseases whose molecular basis is well understood. Thus, a common genetic disease, cystic fibrosis, is due to mutations in a membrane protein. As a result of this mutation, this protein is unable to leave the ER despite the fact that it remains functional. Presumably, this is a case in which the ER quality control machinery acts as a “double-edged sword”: Preventing export from the ER of a slightly misfolded protein causes a much more severe disease due to a complete lack of the protein at its proper location, the plasma membrane. Similarly, α1-antiprotease deficiency, which results in emphysema, is due to a trafficking defect in which the misfolded protein fails to leave the ER.
In most cases, our understanding of acquired disorders, including degenerative diseases, does not extend to the relative importance of disorders of trafficking. However, proteins that undergo complex trafficking pathways, including growth factors and hormone receptors, are central to many of these disorders, suggesting that important connections are yet to be discovered.
An emerging principle of cellular pathophysiology seems to be that for the most important pathways, cells have evolved backup systems that can to some extent maintain crucial functions. Thus, yeasts lacking SRP are sick and have substantial defects in translocation across the ER but survive thanks to non-SRP-mediated backup systems. Thus, patients with I cell disease have correct lysosomal localization of lysosomal enzymes in some tissues even though all tissues lack the GlcNAc phosphotransferase necessary to generate the ligand for the M6P receptor. Two consequences of this principle are that important trafficking defects are apt to be acquired rather than inherited and that identification of these lesions will prove extremely difficult and will require a more intimate understanding of the complex regulation of normal trafficking in the absence of disease.
HORMONE EXPORT NOT MEDIATED BY MEMBRANE VESICLES
Steroid hormones are mainly synthesized in the adrenals, gonads, and placenta. Recent experiments indicate that steroids are synthesized in the nervous system as well, and other tissues may also synthesize steroid hormones in much smaller quantities. The overall pathway for the synthesis of all steroid hormones is similar (Figure 2-5). Tissue-specific, cell-specific, and even subcellular compartment-specific differences in the expression of particular steroidogenic enzymes regulate the particular pattern of steroid hormone synthesized.
All steroids are derived from pregnenolone (Figure 2-5). Pregnenolone, naturally occurring progestins, glucocorticoids, and mineralocorticoids contain 21 carbons and are referred to as C-21 steroids. Androgens and estrogens have two less carbons, and are therefore C-19 steroids. The different rings of the steroid structure are designated A-D, and the numbering of each carbon atom is useful in understanding how the different steroid-synthesizing enzymes modify the steroids at particular locations. All steroids have the basic cy clopentanoperhydrophenanthrene ring (four-ring) structure, and all have a single unsaturated carbon-carbon double bond except for the estrogens, in which the A ring is aromatized. Pregnenolone and other steroids called Δ5 steroids have this double bond between carbons 5 and 6, while progesterone, glucocorticoids, mineralocorticoids, and androgens are called Δ4 steroids and have this double bond between carbons 4 and 5.
Steroids are given both chemical and trivial names. As the chemical names are often cumbersome, trivial names are commonly used. For example, the chemical name for progesterone is pregn-4-ene-3,20-dione, cortisol is 11β,17α,21-trihydroxypregn-4-ene-3,20-dione, etc. Chemical names are given in a particular order: hydroxyl groups, aldehyde groups, core ring structure, and aldehydes or ketones. The core ring structure is
given a name based on the number of carbons it contains. Thus, C-27 steroids are cholestanes, C-21 steroids are pregnanes, C-19 steroids are androstanes, and C-18 steroids are estranes. Like all chemical names, a saturated structure is given the suffix -ane while an unsaturated structure is given the suffix -ene.
Figure 2-5. Pathways of synthesis of the major classes of steroid hormones. Cholesterol is derived from acetate by synthesis or from lipoprotein particles. The numbering of the steroid molecule is shown for pregnenolone. The major pathways thought to be used are shown. (See also Figures 9-4, 12-2, 13-4, and 14-13.)
In addition to the trivial names for the steroid hormones, steroids may be given a letter abbreviation. These are the names of the steroids as they were originally identified by chromatography by the chemists who first isolated and characterized them about 50 years ago. Therefore, cortisol may be called compound F; cortisone, compound E; corticosterone, compound B, etc. The trivial names and systematic names of some natural and synthetic steroids are listed in Table 2-1. (See also Chapters 9, 10, 12, and 13.)
Cholesterol Synthesis & Uptake
All steroid hormones are synthesized from the precursor cholesterol. There are three sources of cholesterol for use in steroid hormone synthesis: de novo synthesis from acetate, pools of cholesteryl esters in steroidogenic tissues, and dietary sources. About 80% of the cholesterol used for steroid hormone synthesis comes from dietary cholesterol, transported in human plasma as low-density lipoprotein (LDL) particles. In rats and other species, cholesterol is transported as high-density lipoprotein (HDL) particles. Uptake of LDL cholesterol is enhanced by ACTH treatment, which increases the activity of LDL receptors and uptake of LDL cholesterol. The majority of LDL cholesterol uptake is by receptor-mediated endocytosis via coated pits, and less
than 10% enters the cell independently of this mechanism. HDL cholesterol uptake, however, occurs by receptor-independent mechanisms.
Table 2-1. Trivial and chemical names of some natural and synthetic steroids.
Intracellular Cholesterol Storage & Transport
Several factors have been identified as being important in cholesterol mobilization within steroidogenic tissues. Cholesterol is esterified to polyunsaturated fatty acids in the endoplasmic reticulum by acyl-CoA:cholesterol acyltransferase (ACAT), and accumulated cholesteryl esters form lipid droplets. These fatty acid esters are hydrolyzed to free cholesterol by cholesteryl esterase (sterol ester hydrolase). Tissue-specific tropic hormones (eg, ACTH, LH, and FSH) stimulate the esterase and inhibit the acyltransferase, resulting in increased accumulation of free cholesterol. Some steroids, such as pregnenolone, testosterone, and estradiol, can be reesterified to form fatty acid esters. Pregnenolone fatty acid esters form biosynthetic intermediates in the synthesis of adrenal steroids, and testosterone and estradiol fatty acid esters can accumulate in target tissues and behave as endogenous long-acting androgens and estrogens.
Steroid Hormone Synthesis
The first step in steroid hormone synthesis, the conversion of cholesterol to pregnenolone, occurs in the mitochondria. However, cholesterol does not enter the mitochondria freely but rather must be carried through the cytoplasm to the inner mitochondrial membrane. Free cholesterol, which would be insoluble free in the aqueous cytoplasm, binds to sterol carrier protein-2 (SCP-2), which plays a major role in its transport to the mitochondria. Once there, cholesterol must traverse the outer mitochondrial membrane and the intermembrane space to reach the inner mitochondrial membrane, where the first steroid hydroxylating enzyme, P450scc, resides. A newly characterized protein, given the name StAR (steroidogenic acute regulator), functions to carry out the rate-limiting step in cholesterol transport across the mitochondrial outer membrane. Since steroids are not stored, their synthesis and its cessation must be exquisitely controlled. A novel mechanism of regulation involving StAR activity has recently come to light. It appears that the nuclear-encoded StAR protein is localized to the mitochondrial matrix—yet it has all of its known enzymatic activity on the outer surface of the outer mitochondrial membrane, a location it maintains only transiently during its import. In this case, compartmentalization is used to provide hair-trigger sensitivity to steroid synthesis through its rapid inactivation by substrate depletion upon StAR localization to the mitochondrial matrix.
Most steroidogenic enzymes are members of the cytochrome P450 group of oxidases. All of these enzymes have molecular masses of about 50 kDa and contain a single heme group. They are called “P450” (“pigment” 450) because they all exhibit a characteristic absorbance at 450 nm upon reduction with carbon monoxide. All the steroidogenic P450s function by the same mechanism. They all reduce atmospheric oxygen with electrons from NADPH. These electrons reach the P450 by one or more protein intermediates. For the mitochondrial P450s, two protein intermediates—adrenodoxin reductase and adrenodoxin—are involved. For the microsomal P450s, one protein intermediate, P450 reductase, is involved (Figure 2-6).
The synthesis of all steroid hormones from cholesterol involves one or more of six distinct cytochromes P450 (Figure 2-6), which are expressed in compartment- and tissue-specific patterns, thereby accounting for the differences in compartment- and tissue-specific steroids (eg, made in mitochondria versus ER and in adrenal versus gonads versus placenta). Many of the steroid hydroxylases have multiple enzymatic activities. Purification of the proteins and cloning of the cDNAs encoding these proteins has rigorously demonstrated that these activities indeed reside within single proteins.
In the brain, the expression of the steroidogenic enzymes is developmentally and regionally regulated, ensuring the regulated synthesis of specific neurosteroids. The enzymes are expressed in both neurons and in neuroglia, suggesting that these two cell types must work together in a coordinated fashion to produce the appropriate neurosteroid. The additional pathways leading to the production of the 3α- and 5α-reduced neuro-steroids are shown in Figure 2-6. Neurosteroids such as allopregnanolone, DHEA, and DHEAS function mainly through modulation of neurotransmitter receptors such as GABAA, NMDA, and sigma receptors rather than by binding to intracellular steroid hormone receptors. By binding to these receptors, they augment GABA- and NMDA-mediated ion flux through the receptor channels. Thus, their actions may be quite rapid. Neurosteroids that augment GABA function, such as 3α- and 5α-reduced derivatives of progesterone (allopregnanolone) are potent anxiolytics. Other functions attributed to neurosteroids include stimulation of axonal growth (DHEA), stimulation of dendritic growth (DHEAS), regulation of myelination (progesterone), regulation of neurotransmitter receptor subunit expression, and neuroprotection (DHEA). In addition, recent studies have proposed a role of neurosteroids in premenstrual
syndrome (withdrawal from allopregnanolone) and in depression (allopregnanolone).
Figure 2-6. Principal steroidogenic pathways in classic endocrine tissues. Other steroids, such as steroid sulfates and lipoidal derivatives, are also produced. The names for each enzyme are shown by each reaction. P450scc, mitochondrial cholesterol side-chain cleavage enzyme, mediates 20α hydroxylation, 22 hydroxylation, and scission of the C20-22 bond; 3β-HSD, a non-P450 enzyme mediates both 3β-hydroxysteroid dehydrogenase and Δ5-Δ4 isomerase activities; P450c21 in the endoplasmic reticulum mediates 21-hydroxylation; P450c11β, mitochondrial 11-hydroxylase in the adrenal fasciculata-reticularis, mediates 11-hydroxylation; P450c11AS, mitochondrial aldosterone synthase in the adrenal glomerulosa, mediates 11-hydroxylation, 18-hydroxylation, and 18-oxidation; P450c11B3, mitochondrial 11,18-hydroxylase in the adrenal fasciculata-reticularis that is expressed in rat adrenals only from day 6 to day 36 after birth; P450c17 in the endoplasmic reticulum mediates both 17α-hydroxylation and scission of the C17,20 bond; 17β-HSD, also called 17-ketosteroid reductase or 17KSR, a non-P450 enzyme of the endoplasmic reticulum, catalyzes a reversible reaction and produces testosterone, which may be converted to estradiol by P450arom in the endoplasmic reticulum, through aromatization of the A ring of the steroid nucleus. 11β-HSD is a non-P450 enzyme that inactivates cortisol in target organs by conversion to cortisone. 5α-Reductase, a membrane-bound non-P450, exists in two forms, type I and type II; 3α-hydroxysteroid dehydrogenase is a non-P450 cytoplasmic enzyme that exists in at least three isoforms in human beings. It catalyzes a reversible reaction favoring the 3α reduced steroid and converts dihydroprogesterone to the potent neurosteroid allopregnanolone. It also converts 5α-dihydrodeoxycorticosterone to allotetrahydrodeoxycorticosterone and dihydrotestoserone to androstanediol. (Z F/R, zonae fasciculata and reticularis; ZG, zona glomerulosa.)
The active form of vitamin D—1,25-(OH)2-cholecalciferol (1,25-[OH]2D3)—is derived from vitamin D3 (cholecalciferol; Chapter 8). Cholecalciferol is obtained either from the diet or from the conversion of 7-dehydrocholesterol (present in the skin) in response to ultraviolet irradiation via a 6,7-cis isomer intermediate (provitamin D). Vitamin D2 (ergocalciferol), which is present in plants and differs from vitamin D3 in having a double bond at C22 and C23, a methyl group at C24, and several features of the A ring of the molecule, also serves as a precursor of active vitamin D products. The ultimate vitamin D products are a mixture of compounds derived from these two compounds and can vary depending on the dietary intake (Chapter 8). Vitamins D2 and D3 are transported, bound to a vitamin D transport protein, to the liver, where the actions of a microsomal cytochrome P450c25 convert them to 25-hydroxy derivatives—25-OH-cholecalciferol (25-[OH]D3) for vitamin D3. 25-(OH)D3 then circulates bound to an α-globulin transport protein; in the proximal tubular cells of the kidney, 25-(OH)D3 is converted to 1,25-(OH)2D3 by the actions of a mitochondrial cytochrome P450c1α (see Chapter 8). This latter step is rate-limiting for overall 1,25-(OH)2D3production and is regulated chiefly by PTH and phosphate ions. States of vitamin D deficiency are best assessed by measuring 25-(OH)D3levels in serum.
Arachidonic acid is the most important and abundant precursor of the various eicosanoids in humans and is rate-limiting for eicosanoid synthesis (Figure 2-7). Arachidonic acid is formed from linoleic acid (18:2n-6; an essential fatty acid) in most cases through desaturation and elongation to homo-γ linoleic acid (20:5n-3) and subsequent desaturation. Whereas eicosanoids are not stored by cells, arachidonic acid precursor stores are present in membrane lipids, from which it is released in response to various stimuli through actions of phospholipases. Phospholipase A2 or both phospholipase C and diglyceride lipase catalyze the cleavage of esterified arachidonic acid from the 2 position of glycerophospholipids in the lipid bilayer of the cell (Figure 2-7). The lipid content of the various cells differs, and this results in different patterns of eicosanoid production from different cell types. Phospholipase A2 activity in vitro can be strongly inhibited by glucocorticoids through the induction of proteins called lipocortins; this may contribute to glucocorticoid suppression of certain inflammatory reactions, but the importance of this block in humans is not established.
Arachidonic acid can be converted to the endoperoxide prostaglandin H2, which is the precursor to the prostaglandins, prostacyclins, and thromboxanes, or it can be acted on by other lipoxygenases to form the leukotrienes and other eicosanoids such as HETE. For prostaglandin synthesis, cyclooxygenase (also called endoperoxide synthetase) converts arachidonic acid to the unstable endoperoxide PGG2, which is rapidly reduced to PGH2. Cyclooxygenase is widely distributed throughout the body (except for erythrocytes and lymphocytes) and is inhibited by aspirin, indomethacin, and other nonsteroidal anti-inflammatory agents. In some tissues, PGH2 can be converted to other prostaglandins (eg, PGD2, PGE2, PGF2 [via PGE2]) in reactions involving prostaglandin synthetases. Similarly, prostacyclin synthetase, which is prevalent in endothelial and smooth muscle cells, fibroblasts, and macrophages, can convert PGH2 to prostacyclins, and thromboxane synthetase, prevalent in platelets and macrophages, can convert PGH2 to thromboxanes (eg, thromboxane A2). Arachidonic acid metabolism by 5-lipoxygenase results in leukotriene production, and metabolism by 12-lipoxygenase results in 12-HPETE (hydroperoxyeicosatetraenoic acid) that is converted to HETE. Arachidonic acid can also be oxygenated by cytochrome P450 monoxygenases to various omega oxidation products and epoxides and derivatives that may have biologic activities.
Historically, it has been believed that prostaglan-dins, like steroids, simply diffused across membranes. Recently, however, a prostaglandin transporter has been identified, suggesting specific uptake of these products by target tissues.
METABOLISM, TRANSPORT, ELIMINATION, & REGULATION OF HORMONES
Hormones circulate both free and bound to plasma proteins. There are major differences between the various hormones in the extent of their association with the plasma proteins. In general, the binding of hormones to plasma is through noncovalent interactions, although cholesterol is considered to be bound through ester bonds to phosphatidylcholine. However, even in this case, the esterified cholesterol is bound through hydrophobic interactions to the lipoprotein particles.
Metabolism of Polypeptide Hormones
The metabolism a polypeptide hormone undergoes depends on which pool it belongs to. As described earlier, some proteolytic events such as signal sequence cleavage occur very early in biogenesis. Other events such as reverse
translocation to the cytoplasm for degradation occur only for a particular subset of chains. Secretory granules full of polypeptide hormone that has been stored for too long undergo fusion to lysosomes and degradation. Finally, after secretion, polypeptide hormones have a distinctive half-life in the bloodstream.
Figure 2-7. Pathways for synthesis of the major classes of eicosanoids: prostaglandins, prostacyclins, thromboxanes, and leukotrienes. All steps in the pathways are not shown. Below each pathway, enclosed by boxes, is a representative compound of the class. (HETE, hydroxyeicosatetraenoic acid; PGG2, prostaglandin G2; PGH2, prostaglandin H2.)
Most polypeptide hormones circulate at low concentrations unbound to other proteins, though there are exceptions. For example, there are several different IGF-I-binding proteins that bind IGF-I. Vasopressin and oxytocin are bound to neurophysins. Growth hormone binds to a protein that is identical to the hormone-binding portion of the growth hormone receptor.
In general, peptide hormones have short half-lives (a few minutes) in the circulation, as seen with ACTH, insulin, glucagon, PTH, and the releasing hormones.
The glycosylated glycoprotein hormones are more stable. Glycosylated chorionic gonadotropin has a half-life of several hours. Although there may be some degradation of the hormone by proteases in the circulation, the major mechanism for hormone degradation is through binding to cell surface receptors for the hormone or through non-receptor cell surface hormone-binding sites, with subsequent uptake into the cell (internalization; see below) and degradation (by enzymes in the cell membrane or inside the cell). A number of specific enzymes mediate these processes, which differ for the various hormones. In addition, several steps may be involved. The first of these may inactivate the hormone, and this can be due—eg, in the case of insulin—to reduction of disulfide bonds in the protein. An important overall source for these enzymes is the lysosome, which may fuse with endocytosed vesicles to expose its enzyme contents and its acid environment to the internalized hormone-receptor complex. An advantage of the short circulating half-lives of some classes of hormones is that the duration of the response can be relatively short. In addition, the persistent presence of significant levels of many classes of hormones that act on the cell surface results in down-regulation of the responsiveness to the hormone that may not be desirable.
Metabolism of Steroid Hormones & Vitamin D
All of the steroid hormones are bound to plasma proteins to some extent, with high-affinity binding to specific globulins and relatively low-affinity and nonspecific binding to proteins such as albumin. The major specific binding proteins are corticosteroid-binding globulin (CBG; transcortin), which binds both cortisol and progesterone; and sex hormone-binding globulin (SHBG), which binds testosterone and estradiol (testosterone more tightly than estradiol). These proteins are present in sufficient concentrations so that over 90% of the total cortisol and about 98% of the testosterone and estradiol are bound. The levels of their binding capacities in some cases exceed only slightly the normal concentrations of the steroid, so that with higher levels a much higher proportion of the hormone can be free. For example, the CBG capacity for cortisol is about 25 ľg/dL (690 nmol/L). Aldosterone does not bind to a specific protein, with the result that only about 50% of the plasma aldosterone is bound. Vitamin D circulates mostly bound to vitamin D-binding protein. This protein binds 25-(OH)D3 more tightly than 1,25-(OH)2D3 or vitamin D3.
The hydrophobic steroid hormones and the D vitamins are filtered by the kidney and generally reabsorbed. About 1% of the cortisol produced daily ends up in the urine. Steroid hormones are ordinarily handled by metabolizing them to inactive species and to water-soluble forms that are more effectively eliminated. It is the free steroid fraction that is accessible to metabolic inactivation. The inactivations are accomplished by converting hydroxyl groups to keto groups, reducing double bonds, and conjugating the steroids with glucuronide and sulfate groups. Over 50 different steroid metabolites have been described.
The production of active hormones by metabolism in peripheral tissues, as is seen with androgens, estrogens, and vitamin D, is discussed above in the section on hormone synthesis. In addition, metabolism in peripheral tissues can determine the type of steroid that binds to the receptor. Aldosterone is ordinarily the major mineralocorticoid hormone responsible for the salt-retaining actions of the steroid hormones. This steroid binds to the mineralocorticoid receptor only about ten times more tightly than cortisol, whose total and free concentrations in the circulation are about 1000 times and 100 times (respectively) those of aldosterone; thus, cortisol should ordinarily be the main occupant of the mineralocorticoid receptors. This, in fact, occurs in tissues such as the brain and pituitary, but in the kidney, cortisol is avidly converted to the essentially inactive metabolite cortisone and perhaps other species which do not bind to the mineralocorticoid receptor. Some types of licorice or genetic defects can block the enzyme responsible for the conversion and result in a mineralocorticoid excess state that is due to cortisol (Chapter 10).
Metabolism of Thyroid Hormones
Thyroid hormones circulate bound to plasma proteins such that 0.04% of the T4 and 0.4% of the T3 are free. About 68% of the T4 and 80% of the T3 are bound by the glycoprotein thyroid hormone-binding globulin (TBG). About 11% of the T4 and 9% of the T3 are bound to transthyretin (thyroid hormone-binding prealbumin; TBPA). The remainder is bound to albumin.
The metabolism of thyroid hormones is discussed in Chapter 7. The circulating half-lives of T4 (7 days) and T3 (about 1 day) are longer than for most hormones. These differences are due to the higher affinity of T4 than T3 for TBG. The hormones are degraded to inactive forms by microsomal deiodinases. The type I 5′-deiodinase is prevalent in most peripheral tissues, including liver and kidney, and is responsible for most of the production of T3. A type II 5′-deiodinase present in the pituitary and central nervous system is involved in generating T3 for feedback inhibition of TSH release. The 5′-deiodinases also convert reverse T3 (3,3′,5′-l-triiodothyronine) to 3,3′-T2 (3,3′-diiodothyronine). 5-Deiodinases act on T4 to generate reverse T3 and on T3 to generate 3,3′-T2. Deaminations and decarboxylations
of the alanine side chains as well as conjugations with glucuronic acid and sulfate groups are also involved in degrading thyroid hormones. (See Chapter 7.)
Metabolism of Catecholamines
The metabolism of the catecholamines is discussed in Chapter 11. These compounds are cleared rapidly, with half-lives of 1-2 minutes. Clearance is primarily by cellular uptake and metabolism, and only about 2-3% of the norepinephrine that enters the circulation is excreted in the urine. Furthermore, a significant amount of the catecholamine metabolites in the circulation reflect catecholamines whose degradation occurred within adrenergic neuron terminals, a point of importance for interpreting clinical data. The catecholamines are degraded by two principal routes, catechol-O-methyltransferase (COMT) and monoamine oxidase (MAO). The measurement of some of the metabolites—normetanephrine, metanephrine, and vanillylmandelic acid (VMA)—can be useful in evaluating cases of possible catecholamine overproduction.
Metabolism of Eicosanoids
Prostaglandins are rapidly metabolized within seconds by enzymes that are widely distributed. Prominent in the metabolism is oxidation of the 15-hydroxyl group of the prostaglandin that renders the molecule inactive. Subsequent other reactions involve both oxidations and reductions.
Regulation of Hormone-Binding Proteins in Plasma
The levels of the plasma-binding proteins can vary with both disease states and drug therapy. For example, CBG, SHBG, and TBG levels are increased by estrogens. SHBG levels are increased by thyroid hormones, and SHBG and TBG levels are decreased by androgens.
An understanding of the roles of plasma binding of hormones is just emerging. In general, the hormones are sufficiently soluble to circulate unassociated at levels at which they are highly active. Cholesterol may be an exception if it is considered a hormone. With the steroid and thyroid hormones, deficiency states characterized by genetic defects with very low levels of transport proteins are not associated with clinical abnormalities. The transthyretin gene, which encodes the protein responsible for most of the thyroid hormone binding in the plasma of mice, has been deleted in this species and the animals were phenotypically normal. Thus, there is no evidence that these binding proteins are essential.
In most cases, it is the free hormone that is active; the free levels of the hormone are responsible for the feedback and related regulatory influences that control hormone release (see below). The free levels of hormones are related to their clearance rates. Clinical states correlate best with the free levels of hormones. The latter is a critical consideration in many cases, as with states of adrenal or thyroid hormone excess and deficiency. With these hormones, factors that affect the levels of plasma binding proteins can spuriously elevate or depress the total hormone levels in otherwise normal individuals, or the changes could mask pathologic hormone excess or deficiency states. These considerations are discussed in Chapters 7, 8, 9. However, transport proteins may greatly facilitate an even delivery of hormones to the target tissues. In a tissue such as the liver, for example, a hormone that is totally free would be completely sequestered as the blood flows through the proximal portions of the tissue, whereas if it were mostly bound, the free hormone would be sequestered in proximal portions and additional hormone would be available for more distal portions through the dissociation of plasma-bound hormone. The latter facilitates more even delivery throughout the tissue. With polypeptide hormones, plasma binding can increase the half-life of the hormone in the circulation; it may also facilitate its delivery into the target tissues.
Overall Regulation of the Endocrine System
The effective concentration of a hormone is determined by the rates of its production, delivery to the target tissue, and degradation. All of these processes are finely regulated to achieve the physiologic level of the hormone. However, the importance of the steps may differ in some cases. By far the most highly regulated process is hormone production. With many classes of hormones, the short half-lives of the hormones provide means of terminating the responses and thus preventing excessive responses. The latter are also blunted by negative regulation of both hormone responsiveness (discussed below) and release, as well as by other factors. For example, in stress, glucocorticoids produced in excess probably blunt the actions of a number of hormones that would otherwise be harmful (Chapter 9). Thus, when the actions and half-lives of the hormones are short, the hormone response can be terminated by simply stopping release of the hormone. An exception is thyroid hormone, with its long half-life. Details of the controls of the individual systems are provided in later chapters.
There are a number of different patterns of regulation of hormone release. Many hormones are linked to the hypothalamic-pituitary axis (discussed in detail in the section on neuroendocrinology; see Chapter 5). These involve both classic feedback loops by hormones that are released by peripheral glands (cortisol, thyroid
hormone, etc) and more subtle control, as is seen with GH and PRL. However, many other systems are more free-standing. This is illustrated by the parathyroid glands (Chapter 8) and by the pancreatic islets (Chapter 17). With the parathyroid glands, the Ca2+ concentration that is increased in the plasma by the hormone exerts a dominant feedback inhibition on the release of PTH. With insulin, depression of glucose levels in response to insulin action results in cessation of the stimulus to release more insulin. In addition, in both cases, the release of the hormone and the overall state of the gland are influenced by numerous other factors.
The stimuli to regulate hormone production include essentially all of the types of regulatory molecules, including hormones such as the tropic hormones and counterregulatory hormones (discussed above), traditional growth factors, eicosanoids, and ions. For example, potassium ion is an important regulator of the adrenal zona glomerulosa. The production of the various eicosanoids is regulated by local factors acting on the cells in which these products are released. For example, tropic stimulation of most endocrine glands results in enhancement of eicosanoid production.
The production of hormones is regulated at multiple levels. First, synthesis of the hormone can be regulated at the level of transcription, as is commonly seen with the polypeptide hormones or the enzymes involved in the synthesis of other hormones such as the steroids. It can also be affected by posttranscriptional mechanisms. Second, release of the hormone stored in secretory granules from tissues that employ the regulated secretory pathway is controlled by secretagogues, as was discussed in the section on hormone synthesis. The secretory cells can store the peptide hormones in sufficient quantity so that the amount released over a short period can exceed the rate of synthesis of the hormone. And third, stimulation of endocrine glands by tropic hormones and other substances such as growth factors can increase the number and size of cells that are actively producing the hormone.
Aridor M, Balch WE: Integration of endoplasmic reticulum signaling in health and disease. Nat Med 1999;5:745.
Bose H, Lingappa VR, Miller WH: Rapid regulation of steroidogenesis by mitochondrial protein import. Nature 2002;417:87.
Compagnone NA, Mellon SH: Neurosteroids: Biosynthesis and function of these novel neuromodulators. Front Neuroendocrinol 2000;21:1.
Gerber SH, Sudhof TC: Molecular determinants of regulated exocytosis. Diabetes 2002;51(Suppl 1):S3.
Gonzalez L, Scheller RH: Regulation of membrane trafficking: structural insights from a rab/effector complex. Cell 1999; 96:755.
Huete-Perez JA et al: Protease trafficking in two primitive eukaryotes is mediated by a prodomain protein motif. J Biol Chem 1999;274:16249.
Huopio H et al: KATP channels and insulin secretion disorders. Am J Physiol Endocrinol 2002;283:E207.
Kanai N et al: Identification and characterization of a prostaglandin transporter. Science 1995;268:866.
Kelly RB: Deconstructing membrane traffic. Trends Cell Biol 1999;9:M29.
Lingappa VR, Farey KC: Physiological Medicine. McGraw-Hill, 2000.
Lingappa VR et al: Conformational control through translocational regulation: a new view of secretory and membrane protein folding. Bioessays 2002;24:741.
Maechler P, Wollheim CB: Mitochondrial glutamate acts as a messenger in glucose-induced insulin exocytosis. Nature 1999; 402:685.
Mellon SH, Griffin LD: Neurosteroids: biochemistry and clinical significance. Trends Endocrinol Metab 2002;13:35.
Miller WL, Mellon SH: Steroidogenesis. In: Reproductive Medicine: Molecular, Cellular and Genetic Fundamentals, 2nd ed. Fauser BCJM (editor). Parthenon Press, 2002.
Miller WL: Molecular biology of steroid hormone synthesis. Endocr Rev 1988;9:295.
Miller WL: Steroid hormone biosynthesis and actions in the materno-feto-placental unit. Clin Perinatol 1998;25:799.
Molinari M, Helenius A: Chaperone selection during glycoprotein translation into the endoplasmic reticulum. Science 2000; 288:331.
Omura T, Morohashi K: Gene regulation of steroidogenesis. J Steroid Biochem Mol Biol 1995;53:19.
Schubert U et al: Rapid degradation of a large fraction of newly synthesized proteins by proteasomes. Nature 2000;404:770.
Smith SS et al: GABAA receptor α4 subunit suppression prevents withdrawal properties of an endogenous steroid. Nature 1998;392;926.
Travers KJ et al: Functional and genomic analyses reveal an es-sential coordination between the unfolded protein response and ER-associated degradation. Cell 2000;101:249.
Warren G, Mellman I: Bulk flow redux? Cell 1999;98:125.