The Isolation and Partial Characterization of Two Novel Sphingolipids from Neurospora crassa:

Neurospora crassa strains labeled uniformly with ^(32)P_i and [^3H]inositol exhibit at least six phospholipid components containing ^3H when separated by paper chromatography. One of the major components is phosphatidylinositol. Other components, which account for 40 to 60% of the lipid-extractable ^3H in various strains, are stable to mild alkaline methanolysis and appear to be sphingolipids with equivalent amounts of inositol and phosphorus. The major phosphosphingolipid was purified by means of differential solubility and by column chromatography on porous silica beads. This substance contains equivalent amounts of hydroxysphinganine and hydroxytetracosanoic acid and 2 eq each of myoinositol, phosphorus, and sodium. Alkaline degradation yielded 2 eq of inositol monophosphate and periodate degradation gave a C-15 fragment. The elemental composition of this compound also fits the formulation, (inositol-P)_2-ceramide. 
A [^3H]inositol pulse-chase experiment carried out with an inositol-requiring mutant in exponential growth shows labeled inositol accumulating in the sphingolipid accompanied by decreased labeling in phosphatidylinositol and the acid-soluble fraction. These changes also occur when the chase is carried out during inositol starvation suggesting that degradation of phosphatidylinositol and formation of sphingolipid occurs in the absence of growth. 
A neutral glycosphingolipid was also obtained as a by-product of the phospholipid purification. This substance is provisionally formulated as the ceramide tetrahexoside: [(gal)_3glu]-N-hydroxytetracosonyl-hydroxysphinganine.

presence of these compounds seemed worthwhile. We chose to first examine strains of Neurospora crassa for the presence of these inositol-containing sphingolipids, not only because of the general biochemical utility of this organism but especially because of the intriguing physiological and morphological effects that occur as a consequence of inositol starvation (3-S).
This paper reports on the presence in N. crassa of a major and novel inositol-containing phospholipid and a novel neutral glycosphingolipid. The mixture was adjusted to pH 8.5 with concentrated NHdOH and refluxed with stirring at 54" for 2 hours. While warm, the mixture was filtered with suction through four layers of cheesecloth.
The filtrate was adjusted to DH 5 with acetic acid and allowed to stand for a week at 5".
The bulk of the supernatant was siphoned off and discarded. of "Pi and [2JH]inositol to ensure uniform labeling of the inositol-containing phospholipids. Extraction of lipids from cells in exponential growth phase was carried out by a procedure that was quite efficient (Fig. 1). Control extraction experiments with cells labeled only with [aH]inositol showed < 1 ye of the counts remaining in the residue. The doubly labeled lipid extract after paper chromatography revealed six distinct azP spots (A to F, Fig. 1) that contained significant aH counts. spot E was judged to be phosphatidylinositol. It had the same Rp as yeast phosphatidylinositol in this system ( Fig. 1) as well as on silica gel thin layer plates developed with Solvent I. Spot E was eluted from the paper (Fig. 1) and subjected to mild alkaline methanolysis (14), rendering all of the azP and aH counts water-soluble.
The labeled material in the aqueous phase was identified as glycerophosphorylinositol by two-dimensional paper chromatography with the solvents previously described (15). This showed a single radioactive spot that was superimposed on the yeast glycerophosphorylinositol which was added as internal reference and detected chemically with periodate.
No significant a2P or aH counts from Spot D became water-soluble by the previously described (2) procedure, and all the radioactivity still migrated at the original Rg on silica gel thin layer plates (Solvent I). Because of the substantial water solubility of Spot A, the following alternate methanolysis procedure was employed.
Equal volumes of the chromatogram eluate and 0.2 N NaOH in methanol were mixed and 32Pi and separated by two-dimensional paper chromatography The lipid extract from doubly labeled cells was obtained and processed precisely as detailed in the legend to Fig. 1. 32P-Labeled spots were located by autoradiography, cut out, and counted. The rest of the paper was cut up, counted, and reported FIG. 1. Two-dimensional chromatography of a lipid extract prepared from cells uniformly labeled with 32Pi and [3H]inositol. Neurospora crassa (straig 89601a) was grown at 26" for 13.5 hours in Fries minimal medium supplemented with myoinositol (20.2 pg per ml) after inoculation with 0.5 X lo6 conidla per ml. "Pi was added to a specific activity of 67.6 pCi per Fmole and [2-3H]myoinositol added to a specific activity of 136 pCi per pmole. A 2-ml aliquot was filtered and washed with unlabeled medium and finally placed in cold 5% trichloroacetic acid. After 2 hours in t,he cold, the cells were washed twice with lo-ml portions of Hz0 and stored at -20" overnight.
The mycelial pellet was suspended in 2 ml of HZO-ethanol-diethylether-pyridine (15: 15:5:1, v/v) and treated at 60" for 15 min. After centrifuging, the supernatant was removed, and an additional 1 ml of extracting solvent was added to the residue and extraction was carried out at 60" for 15 min followed by centrifugation.
The two supernatant extracts were combined and a 0.05.ml aliquot was subjected to two-dimensional chromatography (10) on silica gel-impregnated paper followed by autoradiography as described under "Experimental Procedures." and incubated at 30" for 20 min followed by neutralization with 0.012 volume of glacial acetic acid. A control reaction mixture was instantly neutralized.
The entire reaction mixture was subjected to two-dimensional silica gel thin layer chromatography using Solvent I followed by Solvent II. Autoradiography shows Spot A to be unaffected by this procedure.
Thus, the absence of carboxylic ester bonds in Compounds A and D is indicated. Table I shows the quantitative distribution of radioactivity on the chromatogram shown in Fig. 1. The data show that Spots A to D have the same 3H:32P ratio as Spot E, phosphatidylinositol, indicating strongly that Compounds A to D have equimolar amounts of inositol and phosphorus.
Uniform 32P labeling is assured because Pi is the sole source of P in the medium.
Since the strain employed has a strict requirement for myoinositol, no endogenous synthesis can be anticipated.
The only other factor which could complicate the interpretation of the count ratio as a mole ratio is if the [2-3H]inositol might have been catabolized to other 3H-containing substances. This catabolism appears unlikely since no 3H counts were found in the non-inositol-containing phospholipids, phosphatidylserine, phosphatidylcholine, and phosphatidylethanolamine, which are clustered near the solvent fronts (Fig. 1) and which are reported in Table I as the sum of all other 32P lipids.
Spot F appears to be a polyphosphoinositide because of its distinctly lower 3H:32P ratio; however, it has not been further investigated.
It was considered possible that the lipid pattern exhibited by strain 89601a (Fig. 1 a Spots designated as in Fig. 1. * All the rest of the paper was counted, however, the radioactivity is due primarily to streaking of major spots (Fig. 1). In this mutant, equivalent amounts of Spots A and D were found.
Phosphatidylinositol (E) and the unknown Compound A account for most of the lipid inositol in all these strains.
Since Compound A was a major and unidentified lipid, we set out to isolate sufficient quantities of Compound A for further characterization.
Isolation of Compound A and Glycolipid-We initially assumed that Compound A (Fig. 2) was the substance with the composition, mannose (inositol-P)zceramide recognized in yeast (2) because of their similar RF values on paper chromatography (9, 16). Therefore, we adopted a procedure for its isolation patterned after that employed in the preparation of yeast sphingolipids (2) the majority of the phosphorus elutes in a band well before the major carbohydrate-containing material.
In addition in this concentrate a glycolipid of lesser magnitude elutes in the tail end of the phospholipid peak. It can also be seen that the major glycolipid band is free of phosphorus.
Based on thin layer chromatography, fractions containing Compound A ((IP).&', Fig.  2) and fractions containing the major glycolipid were pooled and obtained in a dry state for further analysis as indicated under "Experimental Procedures." We will refer hereafter to the major glycolipid as simply "glycolipid." The final products each exhibited one spot when detected with rhodamine on silica gel chromatography with either Solvent I (Rp values: A, 0.30; glycolipid, 0.26) or Solvent II (Rp values: A, 0.085; glycolipid, 0.25).
Purified Compound A was mixed with a 32P-labeled crude lipid extract and subjected to two-dimensional chromatography as in Fig. 1. The unlabeled Compound A was detected with rhodamine and corresponded with the 32P zone designated A, detected by autoradiography.
When similar comparisons were attempted by thin layer chromatography (Solvent I) it was noted that 32Plabeled Spot A eluted from paper chromatograms gave two very close spots on thin layer chromatograms in about a 10: 1 ratio. The isolated Compound A corresponded exactly to the upper major a2P-labeled spot in experiments where both were mixed prior to thin layer chromatography (Solvent I).
Chemical Characterization of Compound A-Analysis showed purified Compound A could be formulated as the disodium salt of a sphingolipid composed of 1 eq of ceramide and 2 eq of phosphoinositol (Table III).
Tests for hexosamine and uranic acid were negative.
Thin layer chromatography of the long chain base fraction after acid methanolysis showed it to be composed solely of ninhydrin-positive material at the RF of hydroxysphinganine (phytosphingosine) with a trace of high RF material which was presumably anhydrophytosphingosine, a well known artifact of hydrolysis. Gas-liquid chromatography of the long chain base fraction (Table IV) confirmed these observations and showed that the C-18 base predominated with a trace of the C-20 base evident. Retention times expected for the C-20 base were extrapolated from the data of Carter and Gaver (17).
Thin layer chromatography of the fatty acid methyl ester frac-This procedure involving differential solubility produced a sphingolipid-rich concentrate ("Experimental Procedures") in which better than two-thirds of the phosphorus was in the desired compound and which we estimate was obtained in about 25% yield at this stage as judged by silica gel thin layer chromatography (Solvent I).
At this stage of purification it became evident that the Neurospora Compound A was not equivalent to the yeast compound. The Neurospora compound had a slightly higher Rp on thin layer chromatograms and it appeared to consist of two poorly resolved spots both of which were rhodamine-positive, but only the lower one giving a positive glycolipid test with the orcinol-HzSOI reagent.
The sphingolipid concentrate was further purified on a column of porous silica beads (Porasil).
It can be seen from Fig. 2 that was estimated on the fractions indicated. Silica gel thin layer chromatography was carried out on the fractions indicated (t ) and the results are shown schematically in the inset. Lipids were detected with a rhodamine spray (0) and glycolipids were detected with the orcinol-H&O* reagent (O ). Fractions were pooled as indicated (< > ). (ZP)&, (inositol-P)z-   Treatment of Compound A with periodate followed by reduction with borohydride gave a 64y0 yield of pentadecanol and heptadecanol in the expected ratio (Table IV) showing that the 3and 4-hydroxyl groups of the long chain bases were unsubstituted.
The facile formation of inositol monophosphate by alkaline hydrolysis suggests that phosphodiesterbonds are the likely bonds linking the inositols to the lipid moiety of Compound A and possibly to each other. The alkaline lability of such bonds is probably due to the intermediate formation of cyclic inositol phosphate.
Inositol phosphorylceramide from yeast exhibits a similar alkaline lability (2). It remains to be established which positions in both inositols participate in the phosphodiester bonds. Since the 3-and 4-hydroxyls of the long chain base appear to be unsubstituted, the 1-hydroxyl of the long chain base and the hydroxyl of the fatty acid remain as possible sites for the attachment of the phosphoinositol group(s). If Compound A is like all other complex sphingolipids, one should anticipate substitution at the C-l hydroxyl of the long chain base.
Chemical Characterization of Purified Glycolipicl-Carbohydrate assay was carried out after treating the lipid with anhydrous 1 N HCl in methanol at 80" for 24 hours followed by gas-liquid chromatography of the trimethylsilyl methylglycosides (12). The only sugars observed were galactose (1.93 pmoles per mg) and glucose (0.65 pmole per mg). Direct chemical analysis of total hexose by the phenol-HzS04 method gives a significantly higher value, 2.93 pmoles per mg (Table III).
A sample was also hydrolyzed with 2 N HzS04 for 5 hours at 100". After extraction with petroleum ether the aqueous phase was neutralized with Dowex l-HC03-and chromatographed on cellulose thin layer with butanol-pyridine-Hz0 (6:4:3, v/v). Only spots at the RF values of glucose and galactose were observed after detection with the p-anisidinephthalate reagent (18). A test for hexosamine was negative and a carbazole (19) assay for uranic acid gave a value of 0.3 pmole per mg. Hexoses react in the carbazole test to some extent and the galactose and glucose present could be expected to account for this low apparent value for uranic acid.
Equivalent amounts of long chain base (20) and fatty acid were found, indicating that the glycolipid was a sphingolipid (Table III).
Examination of the appropriate fractions after HCl-methanolysis by thin layer chromatography and gas-liquid chromatography gave the same results as found with Compound A, namely, the principal lipid components appeared to be 2-hydrosytetracosanoic acid and hydroxysphinganine (Tables IV  and V) .
The glycolipid (1 pmole of hexose) was treated for 20 min at 30" in 2 ml of 0.1 N KOH in CHC&-CH30H-HZ0 (16:53:5). After neutralization with 0.2 ml of 1 pi acetic acid, 1 ml of CHCla tion after HCl-methanolysis showed it to be exclusively composed of material at the RF expected for methyl esters of monohydroxy fatty acids. Gas-liquid chromatography of the silylated fatty acid ester fraction confirmed these observations and showed that the major component had the retention time found for the 2hydroxy C-24 acid derivative (Table V).
Elemental analysis (Table III) is consistent with the formulation of Compound A as (inositol phosphate)zNa2ceramide, the ceramide being composed of hydroxytetracosanoic acid and hydroxysphinganine. Treatment with aqueous 1 N KOH at 37" for 15 hours resulted in almost all the phosphorus being converted to a water-soluble organic form. Treatment of this product with phosphatase yielded equivalent amounts of free inositol and Pi (Table VI).
We conclude, therefore, that the product of alkaline hydrolysis is inositol monophosphate accounting for all of t,he phosphorus and inositol present in Compound ii.
Although we have no direct evidence that the fatty acid is in an amide linkage, we can rule out the presence of an ester since Compound A survives the mild alkaline methanolysis procedure that completely deacylates ester-containing lipids. and 1.15 ml of Hz0 were added. All of the original carbohydrate was recovered in the organic phase. Silica gel thin layer chromatography (Solvent I) showed ouly the original rhodamimpositive, orcinol-HzS04-positive spot. This stability toward mild alkaline methanolysis indicates absence of acyl ester groups. We conclude therefore that the fatty acid is amide linked.
Periodate oxidation of the intact lipid followed by borohydride reduction gave 96cyc of the expected yield of pentadecanol plus heptadecanol and in the expected ratio (Table IV). Therefore, the 3-and 4-hydroxyl groups of the long chain bases are unsubstituted.
The sugar and nitrogen values obtained (Table III) are slightly lower than anticipated for a ceramidetetrahexoside and would also fit a ceramidepentahexoside.
However, the gas-liquid chromatography results show a galactose to glucose ratio of exactly 3 : 1, and total hexose estimated chemically (Table III) shows a value reasonably close to 4 moles/1332 g. Since there is no objective evidence for any other sugar, or any other component for that matter, we would conclude that all data best fit a provisional formulation as a ceramidetetrahexoside: Strain 37401 inos-was grown for 14 hours in Fries medium supplemented with myoinositol (20 pg per ml) after inoculation with 0.5 X lo6 conidia per ml. 3zPi was added to a final specific activity of 30 MCI per pmole and [2-3H]myoinositol to a specific activity of 136 &i per pmole. Aliquots (2 ml) were filtered on a Millipore filter and washed rapidly with unlabeled medium; those samples dest,ined to be incubated without inositol were washed with inositol-free medium.
These washed samples were then incubated in 12 ml of fresh unlabeled complete medium, with and without 20 pg per ml of myoinositol.
After 5 hours the medium was removed by filtration and the mycelia was treated with trichloroacetic acid, washed with water and lipid extraction, and chromatography performed as indicated in the legend to Fig. 1. The 3H in each fraction has been calculated on the basis of volume of chase medium. NH O=C-r;H-(CH,,,,JH,

OH
The glycolipid isolated was a fortuitous by-product in the isolation of Compound A and so it is difficult to make a judgement about how much of this substance is present. We have, however, subjected crude lipid estracts from strains 89601a and 37401 to two-dimensional silica gel thin layer chromatography (Solvents I and II) and sprayed the plates with the orcinol-HzS04 reagent.
Many spots are evident with the spot in the region of the isolated glycolipid being the most polar and the second or third most intense.
Turnover of Major

Inositol-containing
Lipids-When growing yeast cells, uniformly labeled with [aH]inositol, were transferred to unlabeled growth medium, phosphatidylinositol lost label, about half of which accumulated in the major yeast sphingolipid, mannose (inositol-P)ceramide (11). We have carried out a similar experiment with an inositol-requiring strain of N. crassu. The chase portion of the experiment was carried out in complete growth medium as well as in an inositol-deficient medium ( Table  VII).
The results show that in a 5-hour chase, substantial decreases in label are seen in the acid-soluble fraction and in phosphatidylinositol.
These decreases are accompanied by increased labeling of sphingolipid and the culture medium.
These changes occur even when growth is limited by inositol deficiency.

DISCUSSION
Although Compound A is a novel substance, the composition, Na2(inositol-P)z-ceramide clearly relates it to the yeast inositol phosphorylceramides that contain one or two phosphoinositol moieties (1, 2). It is also related to the more complex plant lipid, phytoglycolipid, extensively investigated by . This relationship is underlined by the fact that in all these substances are found very long chain (22 to 26 C) hydrosy fatty acids and trihydroxy long chain bases.
Spots B, C, and D (Fig. 1)  terns we have reported here demonstrate a 1 :l ratio of phosphorus and inositol, it seems likely that this group of compounds in Neurospora are monophosphoinositol sphingolipids. Several interesting consequences of inositol starvation in N. crassa have been noted.
The lethal effects of inositol starvation observed with inositol-requiring mutants (3) termed "inositolless death" serves as the basis of a widely used technique of mutant selection (4). It was proposed that autolysis and death ensue when proteases are released from particles enclosed by an inositol lipid-rich membrane that ruptures when cells are starved of inositol (6). Striking morphological changes occur in inositolstarved cells (5). It has also been shown that sugar transport by cultures of an inositol-requiring mutant is rapidly and severely affected by changing the inositol content of the medium (7). The continued breakdown of phosphatidylinositol and synthesis of (inositol-P)nceramide which occurs even when growth is limited by inositol starvation (Table VII) may indicate that the deleterious effects of inositol starvation could not only be due to decreased cellular levels of phosphatidylinositol but also due to aberrantly high levels of the inositol-containing sphingolipids. The recognition of a hitherto unknown array of inositol-containing lipids in Neurospora and the availability of methods for their study should provide an opportunity for the study of their role in membrane structure and function.
It seems worthwhile to examine further the details of metabolism of the inositol-containing phospholipids as related to the various phenomena noted above. Pulse-chase experiments in yeast suggested that phosphatidylinositol can apparently serve as a phosphoinositol precursor of the major yeast phosphosphingolipid, mannose (inositol-P)zceramide (11). The results of a similar experiment carried out with Neurospora (Table VII)  interpretation.
i large part of the iabel that accumulates in I"' the sphingolipid must come from t,he acid-soluble pool either 11. directly or via phosphatidylinositol as intermediate. have not yet been described from these sources. *r