![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Howard Hughes Medical Institute and Institute for Molecular Virology, University of Wisconsin—Madison, Madison, Wisconsin 53706
Received 13 June 2003/ Accepted 15 August 2003
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
9 fatty acid (FA) desaturase, which converts saturated FAs (SFAs) to unsaturated FAs (UFAs). Here we further define the relationships between 1a, membrane lipid composition, and RNA synthesis. We show that 1a expression increases total membrane lipids in wild-type (wt) yeast by 25 to 33%, consistent with recent results indicating that the numerous 1a-induced spherules are enveloped by invaginations of the outer ER membrane. 1a did not alter total membrane lipid composition in wt or ole1w yeast, but the ole1w mutation selectively depleted 18-carbon, monounsaturated (18:1) FA chains and increased 16:0 SFA chains, reducing the UFA-to-SFA ratio from 
2.5 to 1.5. Thus, ole1w inhibition of RNA replication was correlated with decreased levels of UFA, membrane fluidity, and plasticity. The ole1w mutation did not alter 1a-induced membrane synthesis, 1a localization to the perinuclear ER, or colocalization of BMV 2a polymerase, nor did it block spherule formation. Moreover, BMV RNA replication templates were still recovered from cell lysates in a 1a-induced, 1a- and membrane-associated, and nuclease-resistant but detergent-susceptible state consistent with spherules. However, unlike nearby ER membranes, the membranes surrounding spherules in ole1w cells were not distinctively stained with osmium tetroxide, which interacts specifically with UFA double bonds. Thus, in ole1w cells, spherule-associated membranes were locally depleted in UFAs. This localized UFA depletion helps to explain why BMV RNA replication is more sensitive than cell growth to reduced UFA levels. The results imply that 1a preferentially interacts with one or more types of membrane lipids. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One positive-strand RNA virus for which membrane association of RNA replication has been studied is brome mosaic virus (BMV), a representative member of the alphavirus-like superfamily of human, animal, and plant viruses (1). BMV has three genomic RNAs. RNA1 and RNA2 encode proteins 1a and 2a, respectively, which are the only viral proteins required for RNA replication. 1a contains an N-terminal domain with m7G methyltransferase and putative guanylyltransferase activities required for capping viral RNA and a C-terminal RNA helicase-like domain (2). 2a contains a central polymerase domain and an N-terminal domain that interacts with the 1a helicase domain (17). RNA3 encodes a cell-to-cell movement protein (3a) and the coat protein. The 3'-proximal coat protein gene is not translatable from RNA3 but only from a subgenomic mRNA, RNA4, synthesized from negative-strand RNA3 (26).
BMV replicates its genomic RNAs and synthesizes subgenomic mRNA in the budding yeast Saccharomyces cerevisiae (16), duplicating the known features of BMV RNA replication in its natural host plant cells. 1a localizes to the outer nuclear envelope or perinuclear endoplasmic reticulum (ER) membrane and induces this membrane to invaginate into the ER lumen, forming 50- to 70-nm diameter spherular vesicles or spherules (9, 33, 39). 1a also directs 2a polymerase and viral RNA templates to these spherules, which become the sites of viral RNA synthesis (7, 33, 39, 43). The interior of these spherules remains connected to the cytoplasm via a membranous neck contiguous with the ER membrane. Similar membrane spherules are associated with RNA replication by other members of the alphavirus superfamily, nodaviruses, and other positive-strand viruses (references 25 and 39 and references therein). The structure, assembly, and function of BMV spherules have multiple similarities to the replicative cores of retrovirus and double-stranded RNA virus virions (39).
We previously reported the isolation of a yeast mutant that inhibits BMV RNA replication approximately 50-fold due to a mutation in the essential yeast gene OLE1 (21). OLE1 encodes 9 fatty acid (FA) desaturase (Ole1p), an integral ER membrane protein and the only enzyme for unsaturated FA (UFA) synthesis in yeast (29). Ole1p converts saturated palmitic acid (having C16 FA chains with no double bonds; hereafter 16:0) and stearic acid (18:0) into unsaturated palmitoleic (16:1) and oleic (18:1) acids. The BMV-inhibiting allele ole1w has a single amino acid substitution (Y212 to C) in the predicted catalytic domain of Ole1p. This mutation blocks BMV RNA replication prior to negative-strand RNA synthesis but does not inhibit cell growth in the absence or presence of BMV components. OLE1 deletion and medium supplementation experiments showed that all effects of the ole1w mutation on BMV replication are due to reduced UFA levels; i.e., BMV RNA replication depends on UFAs but not directly on Ole1p. The results show that BMV RNA replication is highly sensitive to membrane lipid composition and that manipulation of lipid composition by pharmacological as well as genetic approaches may be a useful antiviral strategy.
To better understand the effects of the ole1w mutation and to obtain further insights into the roles of membranes in RNA replication, we combined biochemical, confocal fluorescence, and electron microscopy (EM) analyses to study membrane lipid composition and 1a-membrane, 1a-2a, and 1a-RNA interactions involved in forming the functional RNA replication complex. Here we show that BMV replication protein 1a stimulates membrane lipid accumulation in the absence of other viral components, that the ole1w mutation blocked RNA replication despite allowing all known 1a functions in replication complex formation, and that ole1w not only globally reduced the cellular UFA-to-SFA ratio but preferentially depleted osmium-reactive UFA levels in 1a-associated perinuclear ER membranes.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
FA analysis. Total FA from 10 optical densities at 600 nm (OD600) of yeast were extracted and were converted to methyl esters as described earlier (27). FA species were separated according to chain length and degree of saturation by gas-liquid chromatography and were identified by retention time. The molar amount of each species was measured by using a flame ionization detector (27).
Cell fractionation and RNase sensitivity assays. Yeast cells were treated with lyticase to remove the cell wall as described earlier (39). The resulting spheroplasts were lysed in YLB buffer (50 mM Tris-Cl [pH 8.0], 2.5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 5 µg of pepstatin/ml, 10 µg of leupeptin/ml, 10 µg of aprotinin/ml, and 10 mM benzamidine) and were centrifuged 5 min at 20,000 x g to yield membrane-enriched pellet and membrane-depleted supernatant fractions. To assay the RNase sensitivity of RNA3 in the pellet fractions, membrane pellets were resuspended in YLB and were divided into three portions. Portions A and B received no initial treatment, while portion C was treated with 0.5% NP-40 for 15 min at 4°C. Portions B and C then were treated with 0.01 U of micrococcal nuclease/µl for 15 min at 30°C (39). RNA was extracted from each fraction and were analyzed by Northern blotting as described earlier (21).
Confocal microscopy. A Bio-Rad 1024 double-channel confocal microscope (at the W. M. Keck Laboratory for Biological Imaging of the University of Wisconsin—Madison) was used to visualize and compare the intracellular sites of accumulation of 1a protein, 2a protein, and sec63p. To detect 1a protein, yeast cells were fixed with formaldehyde, permeabilized with Triton X-100, and immunostained with rabbit anti-1a serum followed by donkey anti-rabbit antibodies conjugated to Texas red as described previously (33). GFP-2a and sec63p-GFP were visualized by their intrinsic fluorescence. GFP was fused to the N terminus of 2a (7) and C terminus of sec63p (J. Kahanab and P. Silver, unpublished results). GFP fusion did not interfere with the normal localization and function of 2a or sec63p-GFP in yeast. Images of the intracellular distribution of 1a (red) and 2a (green) or sec63p (green) within the same optical section (0.5 µm) were acquired sequentially and were digitally superimposed to compare the two distributions. To ensure the reproducibility of the results, each experiment was performed several times, and in each experiment hundreds of cells were examined. Representative results are shown in the figures. The effective resolution of the images is about 100 to 200 nm.
EM. For the experiment whose results are shown in Fig. 5, yeast cells were fixed in 4% paraformaldehyde and 2% glutaraldehyde, postfixed with 1% osmium tetroxide and 1% potassium ferricyanide, stained with 1% uranyl acetate, dehydrated in a graded series of ethanol solutions, and embedded in Spurr's resin (Electron Microscopy Sciences, Ft. Washington, Pa.) as described earlier (39). Seventy-nanometer sections were cut and placed on copper grids, poststained with 8% uranyl acetate in 50% methanol and Reynold's lead citrate, and analyzed with a Philips CM120 transmission electron microscope at the Medical School Electron Microscope Facility of the University of Wisconsin. For immunogold EM experiments (see Fig. 6), yeast cells were similarly fixed except that the preembedding osmium tetroxide and uranyl acetate steps were omitted, samples were embedded in LR White resin (Polysciences, Inc., Warrington, Pa.) and sections were placed on nickel grids. Grids were blocked with 0.5% gelatin, immunostained with anti-1a rabbit serum and 12-nm gold-labeled secondary antibodies, poststained, and analyzed by transmission EM as described above.
|
|
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In S. cerevisiae, palmitic acid (16:0), stearic acid (18:0), palmitoleic acid (16:1), and oleic acid (18:1) comprise well over 90% of total FA (14). As shown in Fig. 1A, comparing these major FA species revealed that the ole1w mutation had little effect on 16:1 or 18:0 levels but caused an approximately 25% drop in 18:1 and an 50% rise in 16:0. The selective decrease in 18:1 but not 16:1 FA levels suggests that the ole1w mutation preferentially inhibits action of the encoded desaturase on 18:0 substrates. Similar differential effects of mutations on desaturase specificity for 16:0 and 18:0 substrates have been reported before (46). The associated increase of 16:0 levels may be a secondary consequence of inhibiting 18:1 synthesis, since 16:0 is a substrate for synthesizing 18:0 and thus 18:1.
|
25%. However, due to the simultaneous increase in 16:0, the ratio of total UFA to SFA, which strongly influences membrane fluidity and plasticity, was reduced from approximately 2.5 in wt yeast to 1.5 in ole1w yeast (Fig. 1B). Moreover, further results below imply that variations in local as well as total FA composition may be important in the effects of the ole1w mutation on BMV RNA replication complexes.
In contrast to the effects of the ole1w mutation, expression of 1a had little or no effect on FA composition in either wt or ole1w yeast (Fig. 1A and B). Thus, 1a did not directly or indirectly regulate the activity of the OLE1-encoded 9 FA desaturase.
1a expression increases yeast FA levels. To compare total FA levels per cell, the molar amounts of various FA for each culture were summed, normalized to the OD600 of the culture, and expressed as a percentage of the total FA value for wt yeast lacking 1a (Fig. 1C). To determine whether any differences in total FA content might relate to ole1w- or 1a-induced changes in cell size and whether OD600 accurately reflected cell density, we examined samples of each culture by light microscopy. In addition to visual examination, the cross-sectional areas of 150 to 300 cells from each culture were measured digitally (IPLab Spectrum program; Scanalytics, Inc. Fairfax, Va.). As shown in Table 1, there were no significant differences in cell size for wt or ole1w yeast with or without 1a.
|
In the absence of 1a, the ole1w mutation reduced total FA per cell by 16% relative to wt yeast (Fig. 1C). The basis for this reduction is not yet clear. It appeared tempting to speculate that total membrane synthesis in ole1w mutant yeast might be limited by the mutant's reduced capacity for UFA synthesis and the need to maintain the UFA-to-SFA ratio within viable limits of membrane fluidity. However, in ole1w yeast, 1a expression increased total FA accumulation by 33% (Fig. 1C). The size of this 1a-induced increase in total FA levels, which was even greater than that for wt yeast in both relative and absolute terms, implied that the ole1w mutation did not inhibit BMV RNA replication by preventing 1a-induced synthesis of additional membrane to support formation of the membrane-enveloped spherular replication complexes.
Reduced UFA levels do not alter ER localization of 1a. In wt yeast and in BMV's natural plant host cells, 1a localizes to ER membranes (33, 34). In lysates of ole1w yeast, we previously found that 1a cofractionates with total cellular membranes (21). Thus, the ole1w mutation does not prevent membrane association of 1a. To determine whether 1a localized to its normal ER membrane sites in ole1w yeast, we used confocal microscopy to compare the intracellular distributions of 1a and a yeast ER marker, sec63p, in wt and ole1w yeast. In over 95% of ole1w cells (Fig. 2B), 1a colocalized with sec63p in a pattern indistinguishable from that of 1a in wt yeast (Fig. 2A). In both cases, 1a colocalized with sec63p on perinuclear ER membranes and, to a lesser degree, on peripheral ER membrane strands, which in yeast are frequently appressed to the outer edge of the cell.
|
Reduced UFA levels do not alter 1a-induced ER colocalization of 2a. In ole1w yeast, the 1a functions tested in Fig. 2 and in prior work (21) appeared normal, but viral positive-strand, negative-strand, and subgenomic RNA synthesis was blocked (21). Since viral RNA synthesis requires not only 1a but also BMV 2a polymerase, the possible effects of the ole1w mutation on 2a, its interactions with 1a, or its interactions with cellular components were of particular interest.
When expressed without 1a in wt yeast, BMV 2a polymerase has been found distributed throughout cells (7). However, in the presence of 1a, an N-terminal segment of 2a interacts with the helicase-like region of 1a and colocalizes on ER membranes (7). We showed previously that, in ole1w yeast expressing 1a and 2a, 2a cofractionates with total cell membranes (21). To determine whether 2a colocalized normally with 1a to ER membranes in ole1w yeast, or if 2a might alter the distribution of 1a in such cells, we used confocal microscopy to compare the intracellular distributions of 1a and 2a. Because 2a normally accumulates to low levels relative to 1a and gives fainter signals by immunofluorescence, these experiments used a previously studied, functional GFP-2a fusion that supports BMV RNA replication (7). In the absence of 1a, this GFP-2a was distributed throughout the cell in wt and ole1w yeast (reference 7 and results not shown). When 1a and GFP-2a were coexpressed in ole1w yeast, they were found colocalized in ER-associated patterns indistinguishable from those in wt yeast (Fig. 3). In both wt and ole1w yeast, as observed previously, 1a and 2a signals were particularly strong on perinuclear ER but also extended to peripheral ER strands. Therefore, reduced UFA levels did not detectably perturb 1a-directed 2a localization to the ER, nor did they cause 2a to alter the intracellular distribution of 1a. As in Fig. 2C, a small percentage of ole1w yeast cells showed 1a and 2a colocalization in relatively large, amorphous bodies (results not shown).
|
To determine if the ole1w mutation and associated changes in membrane lipid composition affected the formation, stability, or membrane association of this protected RNA pool, we tested lysates of wt and ole1w yeast expressing 1a and BMV RNA3. 1a-dependent membrane association of RNA3 can be assayed by fractionating such lysates into a membrane-enriched pellet and a membrane-depleted supernatant (39). As shown in Fig. 4A, the degree of 1a-induced membrane association of RNA3 in ole1w yeast was as high as that in wt yeast. Moreover, as shown in Fig. 4B, the membrane-associated RNA3 from ole1w and wt yeast was equally protected from micrococcal nuclease but was equally susceptible to nonionic detergent NP-40 plus micrococcal nuclease. Thus, the ole1w mutation did not detectably alter 1a-dependent membrane association, nuclease protection, or nonionic detergent susceptibility of RNA3.
|
Figure 5 shows representative electron micrographs of wt and ole1w yeast expressing 1a. In wt yeast, the lipid bilayers of the inner and outer perinuclear ER membranes and the invaginated spherule membranes were marked by the typical paired lines of osmium staining (Fig. 5A and 5C). In ole1w yeast (Fig. 5B and D), most sections of the general, perinuclear ER membranes also were marked by osmium staining. However, in these yeast mutants, the lipid bilayer-associated lines of OsO2 deposition often were not as distinctive and clear as those in wt yeast and occasionally were disrupted by short gaps. These results appear consistent with the dependence of osmium staining on adjacent UFA chains and the reduced membrane UFA levels in these cells. This partial, general decline in membrane definition was seen in the peripheral ER and other membranes in ole1w yeast with or without 1a but was slightly more pronounced in the presence of 1a.
In addition, even more notable effects of the ole1w mutation were found in 1a-expressing yeast in specific subdomains of the perinuclear ER membrane (Fig. 5B and D). In ole1w cells expressing 1a, portions of the nuclear border were marked by 50- to 60-nm spherular cores similar to those found in the ER lumen of wt yeast expressing 1a (Fig. 5A and C). However, in the ole1w yeast these regions lacked well-defined osmium staining of either the inner and outer nuclear envelope membrane or of the spherule-bounding membranes surrounding such cores in wt yeast. Immediately flanking regions of the nuclear envelope, however, showed visible osmium staining (Fig. 5B and D). Thus, reactivity of the nuclear envelope membrane with the UFA-dependent reagent OsO4 was significantly reduced in the immediate vicinity of these cores.
To determine whether the 1a-induced, perinuclear, spherule core-like structures in ole1w yeast were counterparts of spherule cores in wt yeast, we used immunogold EM with anti-1a antisera to determine if these structures contained 1a. As is required to preserve 1a antigenicity, for this analysis yeast was fixed only with aldehydes (39). Without OsO4 fixation, membrane lipids are depleted during later alcohol dehydration steps of sample processing for thin sectioning. However, an outline of the membrane normally is preserved due to aldehyde cross-linking of membrane proteins, which comprise about 50% of the total mass of yeast membranes. Consequently, in the resulting electron micrographs, the aldehyde-cross-linked, lipid-depleted membranes generally appear as white, electron-lucent strips, rather than as osmium-stained black lines.
In wt yeast, as seen previously (39), 1a localized primarily to the 50- to 70-nm spherical cores that were found between the inner and outer perinuclear ER membranes (Fig. 6A). Within this ER lumen in wt yeast, these cores usually were separated by additional electron-lucent layers that appeared to correspond to the spherule-bounding membranes seen in Fig. 5A and B. In ole1w yeast (Fig. 6B and C), 1a similarly localized primarily to the dilated, 50- to 70-nm space between the inner and outer nuclear envelopes. However, while close inspection of the micrographs revealed that this dilated ER lumen contained individual spherule cores as in Fig. 5B and D, the separation between these cores was less defined than in wt yeast. Thus, in contrast to Fig. 5, this osmium-independent processing allowed visualization of the expected inner and outer nuclear membranes bordering the 1a-induced spherule cores in ole1w yeast but further illustrated that the membranes flanking spherules in ole1w yeast differed in structure, composition, or both from those in wt yeast.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
9 FA desaturase was caused by much smaller shifts in lipid balance than in previously isolated OLE1 mutants affecting cellular processes (Fig. 1). Thus, viral RNA synthesis was highly sensitive to membrane lipid composition. Moreover, although blocked in even the earliest forms of viral RNA synthesis, ole1w yeast supported all known steps in 1a interaction with ER membranes, BMV 2a polymerase, and viral RNA templates for replication complex assembly, including forming 1a-containing perinuclear spherule cores (Fig. 2 to 6). However, in ole1w yeast, 1a-associated perinuclear membranes were locally deficient in reactivity with the UFA-specific stain OsO4, implying that 1a caused local variations in membrane lipid composition and that the already reduced UFA levels in the mutant yeast were preferentially further depleted in the vicinity of 1a. The relation of these results to other viral and cellular processes is discussed further below. 1a stimulates membrane lipid accumulation. As noted in the introduction, RNA replication by positive-strand RNA viruses occurs on intracellular membranes, often in association with membrane proliferation and vesiculation that, for some viruses, has been shown to be functionally important for RNA replication (5, 12, 24, 30). Here we showed that expressing BMV RNA replication factor 1a alone in yeast increased total membrane lipid accumulation per cell by 25 to 33% (Fig. 1C). The absence of any corresponding increase in cell size or plasma membrane area (Table 1) shows that this must reflect an increase in intracellular membranes, consistent with recent observations that 1a induces vesicle-like invaginations that greatly increase the area of the affected ER membrane and form the compartment in which RNA replication occurs (39). Besides the production of these perinuclear ER spherules, 1a expression did not induce any discernible proliferation of intracellular membranes. 1a may stimulate lipid synthesis indirectly, since 1a accumulation on the ER membrane and possible 1a exclusion of some cell membrane proteins may trigger pathways that normally induce lipid synthesis for maintaining proper lipid-to-protein ratios in cellular membranes. Alternatively, we cannot rule out that 1a directly stimulates lipid synthesis by interacting with one or more of the ER membrane-associated enzymes involved in FA and lipid synthesis.
Globally and locally altered membrane lipid balance in ole1w yeast. In ole1w yeast, BMV RNA replication was blocked by relatively modest shifts in membrane lipid composition (Fig. 1), underscoring that viral RNA replication was much more dependent on lipid composition than was cell growth. Temperature-sensitive (ts) mutations that more severely inhibit OLE1 expression or activity block cell growth and disrupt the morphology of multiple cellular membranes. When shifted to the nonpermissive temperature of 37°C, a yeast strain with ts mutations in MGA2 and SPT23, encoding transcription factors required for OLE1 expression, decreased OLE1 mRNA by 15-fold, total UFA levels by 35%, and the UFA-to-SFA ratio by 2.5-fold; blocked yeast growth; and caused separation of the inner and outer nuclear membranes (48). Shifting mdm2, a ts allele of OLE1, to 37°C decreased UFA levels by threefold and the UFA-to-SFA ratio by over 10-fold, abolished transfer of mitochondria from mother cells to budding daughter cells, and caused the remaining mitochondria to fragment (41). Both strains were rescued by UFA feeding at nonpermissive temperature, confirming that their defects were caused by depleting UFAs.
In contrast, the ole1w mutation blocked BMV RNA replication without detectable defects in yeast growth or morphology. ole1w yeast had normal doubling times, normal cell sizes (Table 1), and no observable abnormality in the morphology of subcellular structures in the presence or absence of BMV components. This lack of cell growth or morphology defects is consistent with the finding that wt yeast maintains UFA levels well above the minimum required for growth under optimal conditions, providing extra membrane fluidity to cope with changes in temperature or other environmental conditions (4). Moreover, the ole1w mutation caused little or no change in the levels of total FA or of the more abundant 16:1 UFA and only a 25 to 30% decrease in the less abundant 18:1 UFA (Fig. 1A and C). Overall, total UFA levels decreased by only 12%, and the UFA-to-SFA ratio fell by only 40%.
Nevertheless, the 12% decrease in UFA levels in ole1w yeast inhibits BMV RNA replication by 95% or more (21). The ole1w mutation did not inhibit 1a-induced membrane synthesis (Fig. 1C), 1a-induced transfer of viral RNA to a membrane-associated, nonionic detergent-susceptible, nuclease-protected compartment (Fig. 4), or formation of 1a-containing, perinuclear membrane-associated spherule cores (Fig. 5 and 6). In ole1w yeast, the normal 1a stimulation of membrane synthesis (Fig. 1C) and lumenal location of 1a-containing cores (Fig. 5B and 6B) suggest that formation of the 1a-containing spherule cores involved ER membrane invagination, as in wt yeast. However, as shown in Fig. 5B, osmium staining of the membranes around the spherule cores was specifically reduced relative to flanking sections of the perinuclear ER membranes. Since such staining depends on osmium reaction with the double bonds of UFA chains in membrane lipids, this implies that the regions of the perinuclear ER membranes containing 1a and 1a-induced spherules were preferentially depleted in UFA-containing lipids.
Since osmium staining involves diester formation between the double bonds of adjacent lipids with UFA chains, the localized reduction in staining may not represent complete UFA exclusion but only depletion below a threshold at which adjacent UFAs become rare. Consistent with this, 1a-containing membranes might have a lower-than-average affinity for UFA-containing lipids even in wt yeast, but the reduction in their UFA concentration might reach an osmium-discernible threshold only when the UFA-to-SFA ratio drops sufficiently below wt, as in ole1w yeast.
The ability of 1a to locally modulate membrane lipid composition is understandable, as 1a is present on the membrane at high local concentration. 1a is one of the most abundant proteins in total nuclear membrane preparations (39), interacts with itself (28), accumulates in localized patches on ER membranes (33, 34), and is present in hundreds of copies per spherule (39).
Any of several mechanisms might contribute to localized depletion of osmium-stainable UFA chains from 1a-containing, spherule-associated membranes. For example, self-interaction (28) and multimerization (39) of membrane-associated 1a proteins might create a membrane zone with reduced protein and lipid mobility. Since the rigid bends of the cis double bonds in UFAs favor a less densely packed, more fluid state, UFA-containing lipids might tend to be excluded from this relatively static membrane zone into the more fluid surrounding membrane, leaving behind a higher concentration of more tightly packing, less fluid, SFA-containing lipids. Alternatively, 1a might selectively interact with and retain SFA-containing lipids, so that the high local 1a concentration would tend to exclude UFAs. Finally, 1a might preferentially interact with UFA-containing lipids but in a way that disrupted their ability to stain with osmium. 1a is a peripheral membrane protein on the cytoplasmic face of the ER but possesses a membrane affinity that is unusually high for a peripheral membrane protein (9). Some peripheral membrane proteins, including cytochrome c, appear to achieve high-affinity membrane interaction by using a hydrophobic protein cavity to bind UFAs on membrane lipids (18, 45). Such binding requires the UFA to extend out of the membrane via rotation around the C2
C3 carbon bond of the phospholipid glycerol backbone. Similar interaction of 1a with UFA-containing lipids might sequester UFAs out of the lipid bilayer and prevent their effective interaction with OsO4 and other lipids, thus further lowering the already reduced membrane UFA levels in ole1w yeast and leading to the locally poor membrane fixation and staining.
Blocked RNA replication in ole1w yeast. The ole1w mutation did not inhibit 1a and 2a colocalization on ER membranes (Fig. 2 and 3), formation of ER-lumenal, 1a-containing spherule cores (Fig. 5 and 6), or 1a transfer of BMV RNA to a membrane-associated, nuclease-protected compartment (Fig. 4). Nevertheless, the ole1w-induced shift in lipid balance, which is particularly pronounced in 1a-associated membranes, induces a severe, UFA-suppressible block to viral RNA replication at or before synthesis of negative-strand RNA (21). This strong sensitivity implies that the membrane is an essential, functional component of the RNA replication complex. While the results presented here define a narrow interval between unaffected and affected steps in replication complex assembly and function, the precise nature of the replication block remains to be determined. Fluid membranes and specific lipids are required to activate some protein functions and to facilitate formation or modulation of protein-protein interactions (18, 19). For example, UFA-containing lipids activate the membrane-associated Escherichia coli DNA replication initiator protein DnaA (8). Similarly, membrane fluidity, UFA-containing lipids, or both may be needed for 1a or 2a polymerase to function properly or to interact properly with each other or host factors.
In addition, local variations in lipid composition have central roles in regulating the formation and properties of curved membranes. Specific lipids are crucial for stabilizing the highly curved membrane junction between nuclear pores and the surrounding nuclear envelope (38). Formation and budding of synaptic vesicles require specific lipid modifications that convert inverted cone-shaped lipids to cone-shaped lipids, thereby favoring a specific polarity of membrane curvature (37). Similarly, local variations in lipid composition likely are crucial for the formation and properties of the vesicular, spherular membrane invaginations that envelope the replication complexes of BMV, nodaviruses, and many other positive-strand RNA viruses (references 25 and 39 and references therein). Since the bend introduced by a cis double bond gives UFA-containing lipids a cone shape promoting negative membrane curvature, the OLE1-dependent UFA-to-SFA ratio is an important contributor to the properties of such curved membranes. Just as for budding synaptic vesicles, lipid composition might be particularly vital for the properties of the highly curved membrane necks joining the light bulb-shaped spherules to the outer perinuclear membranes. These necks appear crucial for RNA synthesis as the likely channels for ribonucleotide import and product RNA export (39).
|
ACKNOWLEDGMENTS |
|---|
This work was supported by the National Institutes of Health through grant GM35072. P.A. is an Investigator of the Howard Hughes Medical Institute.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
9 fatty acid desaturase inhibits brome mosaic virus RNA replication between template recognition and RNA synthesis. J. Virol. 75:2097-2106.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| J. Bacteriol. | Mol. Cell. Biol. | Microbiol. Mol. Biol. Rev. |
|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|
