Novel surface lipids of diapausing Manduca sexta pupae. Long chain oxoalcohol esters of acetoacetic, hydroxybutyric, and acetic acids.

Ester components in the surface wax from diapausing tobacco hornworm pupae, Manduca sexta L., were separated by thin layer chromatography and gas-liquid chromatography, and characterized by infrared spectroscopy and gas-liquid chromatography-mass spectrometry. Three groups of esters were identified as natural derivatives of acetic acid, acetoacetic acid, and 3-hydroxybutyric acid. The major ester fraction was identified as a mixture of C26 (10%), C27 (5%), and C28 (85%) oxoalcohol esters of acetoacetic acid. The major homolog consisted of equal amounts of 11-oxooctacosanyl 3-oxobutanoate and 12-oxooctacosanyl 3-oxobutanoate. Lesser amounts of 11- and 12-oxooctacosanyl and n-octacosanyl esters of acetic and 3-hydroxybutyric acids were also identified. The chain length distributions of these C26, C27, and C28 oxoalcohol and n-primary alcohol ester moieties, as well as the isomeric ratios for the 11- and 12-oxoalcohol isomers, were similar to the oxoaldehydes and unesterified oxoalcohols previously identified by Buckner et al (Buckner, J. S., Nelson, D. R., Haak, H., and Pomonis, J. G. (1984) J. Biol. Chem. 259, 8452-8470) as lipid components of the surface wax of M. sexta pupae.

The major constituents of the surface lipids of the tobacco hornworm, Manduca sexta, have been identified as long chain oxoaldehydes and oxoalcohols (1). These aldehydes and alcohols consisted of a homologous series of c26,1 CZ7, and c 2 S components, with the major component (CZs, 75-85%) consisting of nearly equal amounts of the 11-oxo and 12-oxo isomers. Alkaline hydrolysis of the surface lipids indicated that the majority of the oxoalcohols existed as esters of unknown acids. These acids appeared to be short chain, volatile compounds that were not the usual fatty acids.
To our knowledge, long chain oxoalcohols esterified to short chain acids have not been reported from insects. However, esters consisting of long chain oxoalcohols and long chain oxoacids have been reported as major components of the * A preliminary report was given at the 1983 National Meeting of the Entomological Society of America, November 28-December 2, Detroit, MI. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 'The subscript for lipid components refers to total number of carbon atoms in the lipid. Other abbreviations used are: TLC, thin layer chromatography; GLC, gas-liquid chromatography; GLC-MS, gas-liquid chromatography-mass spectrometry; EI-and CI-MS, electron impact-and chemical ionization-mass spectrometry; NaBH,, sodium borohydride; BSA, N,O-bis(trimethylsily1)acetamide; Me&, trimethylsilyl; TBDMSTFA, N-methyl-N-(t-butyldimethylsily1)trifluoroacetamide; t-BuMe2Si, t-butyldimethylsilyl. surface wax of several insects. A Cm wax ester, 15-oxotetratriacontanyl 13-oxodotriacontanoate, was produced by the cochineal insect, Coccus cacti (2), and the wooly alder aphid, Prociphilus tessatatus (3). Nearly 100% of the surface wax of another cochineal insect, Dactylopius confusus, was the CM oxo ester, 15-oxotetratriacontanyl 11-oxotriacontanoate (3). The rhizomes of the plant Cryptocoryne spiralis contained two oxoacid esters, ethyl 14-oxotetracosanoate and 15-oxoei-cosanyl14-oxoheptadecanoate (4).
In this paper, we present evidence for the composition and identification of novel esters as major components of the surface lipids of M. sexta pupae in diapause. Structures of intact ester components as well as their alcohol and acid moieties were determined by TLC, GLC, IR spectroscopy and GLC-MS.

MATERIALS AND METHODS AND RESULTS~
The surface lipids of M. sexta were separated by TLC into nine fractions, designated Fractions I-IX (1). Fraction I was identified as hydrocarbons (5). Fractions V and VI11 were identified as c 2 6 " c 2 8 oxoaldehydes and C26-cZS oxoalcohols, respectively. The other major lipids, Fractions IV, VI, VII, and IX, were not identified, but they were shown to contain c Z 6 -c Z s oxoalcohols and n-primary alcohols esterified to unknown short chain acids.

Identification of Fraction ZV: Acetate Esters
Fraction IV migrated just ahead of the oxoaldehydes (Fraction V) when chromatographed on silica gel plates developed with solvent system A ( Fig. 1). GLC analysis of Fraction IV showed three components designated as peaks 1-3 (Fig. 2). The major component of Fraction IV (peak 3) had an equivalent chain length of 33.7. In comparison, the CZ8 oxoaldehyde in Fraction V had a value of 31.8. The column retention times of GLC peaks 1-3 relative to the retention times of alkane standards suggested that Fraction IV was a homologous series of components.
The major component of Fraction IV (Fig. 2, peak 3) was obtained pure by trapping from the gas chromatograph. Analysis by IR spectroscopy indicated two intense absorption bands at 1700 and 1740 cm", indicative of the carbonyl for aliphatic ketones and esters, respectively (Fig. 3A) homolog was primarily the 11-0x0 isomer, whereas the C27 homolog was a mixture of 11-0x0 and 120x0 isomers.
ponent of Fraction IV was analyzed also by EI-and CI-mass spectrometry and was concluded to be a mixture of the acetate esters of ll-and 12-oxooctacosanol (Fig. 4). The CI mass spectrum (Fig. 4B) indicated a molecular weight of 466 as a result of an M + 1 ion at m/z 467 (base peak) with isobutane as ionizing gas. The ion at m/z 407 corresponds to the loss of acetate. a-cleavage at the 12-0~0 position was indicated in the CI mass spectrum as well as the EI mass spectrum ( Alkaline hydrolysis of Fraction IV yielded the same homologous series of C26, C&, and C,, oxoalcohols that were previously found in Fraction VIII (1). Also, the GLC-EI mass spectrum observed for the major component of Fraction IV (Fig. 4A), was the same as the mass spectrum obtained for the acetate derivative of the Cga oxoalcohol derived from acetylation of Fraction VIII. In addition to the acetate ester of the C& oxoalcohol, the minor peaks 1 and 2 ( Fig. 2) were identified from their GLC-EI mass spectra as the acetate esters of Czfi and Cz7 oxoalcohols (data not shown). The Cl6 Identification of Fraction VI: Acetoacetate Esters Characterization of Fraction VI and Its Hydrolysis Products--Alkaline hydrolysis of surface lipid extracts from M. sexta pupae indicated that Fraction VI contained esters of oxoalcohols (1). This observation was confirmed by TLC separation of Fraction VI and subsequent hydrolysis (60 "C for 2 h in sealed ampules with 0.2 M KOH in 80% isopropanol) to yield oxoalcohols. Me:,Si derivatives of the oxoalcohols were identified by GLC-MS as being the same homologous series of compounds that had been previously identified for the oxoalcohols and for the partially reduced oxoaldehydes, namely, C,, (5-lo%), Cpi (2-5%), and Cgx (8590%). The C,, oxoalcohols of Fraction VI contained mainly the 12-0~0 isomer with a lesser amount of the 11-0x0 isomer. Consistent with earlier findings for the oxoalcohols and oxoaldehydes, the 11-0x0 isomer was predominant for the C,, homolog.
IR spectroscopic analysis of purified Fraction VI also indicated an ester; a characteristic absorption band at 1740 cm-' (Fig. 3C) that had been observed for the acetate ester of oxoalcohols ( Fig. 3A) and the band at 1700 cm-' characteristic of aliphatic ketones. The acetate esters of oxoalcohols (Fraction VI) had a greater RF on TLC plates than that of Fraction VI. Hence, the acid moiety of Fraction VI must be more polar than acetate.
Fraction VI possessed poor chromatographic properties when analyzed by GLC (Fig, 5A). Only one major peak was observed having a broad asymmetrical shape. Mass spectral scans taken across this peak showed a molecular ion at m/z 424, a fragment ion at m/z 253, and rearrangement ions at m/z 214 and 268 and indicated that the peak was mainly 12- oxooctacosanol (Fig. 6). Fragment ions in the mass spectrum also indicated the presence of lesser quantities of the 11-oxo isomer. Therefore, in contrast to the acetate esters, Fraction VI apparently underwent hydrolysis during GLC-MS analysis and only the oxoalcohol moieties were detected.
More useful mass spectra for Fraction VI were obtained from solid-sample-probe analysis in both the E1 and CI modes. The ions at m/z 423 and 425 in the E1 and CI mass spectra, respectively, indicated the presence of the CzR oxoalcohol (Mr = 424) hydrolysis product (Fig. 7, A and B). The M + 1 ion at m/z 509 in the CI mass spectrum indicated a molecular weight of 508 for the intact ester of Fraction VI. Structures consistent with these E1 and CI mass spectra were the 3oxobutanoate (acetoacetate) esters of 11-and 12-oxooctacosanol. The fragment ion at m/z 103 suggests the protonated 3oxobutanoic acid, CH3-C(0)-CH2-C(O)OHt as shown for other fatty alcohol esters (7,8). Other fragment ions that support the structural assignment Esters of acetoacetic acid can exist as tautomers, both in the keto and enol form. The enol isomer reacts with ferric chloride to give a red or violet color (9). TLC plates spotted with either total lipid or purified Fraction VI and developed with either solvent system A or B were sprayed with a solution of ferric chloride. With mild heating and usually within 5 min, a reddish spot appeared at a position on TLC plates coincident with Fraction VI. A ferric chloride reaction was not detected with any of the other surface lipid components. However, the same coloration was observed with either ethyl or butyl acetoacetate standards spotted on TLC plates. Enolization of Fraction VI was also indicated from IR spectroscopy by a band at 1645 cm", characteristic of a shift to a shorter carbonyl frequency due to enol tautomerization of poxo esters (10).
NaBH4 Reduction of Fraction VI-If Fraction VI was a mixture of 11-and 12-oxoalcohol esters of acetoacetic acid, the %oxo of the acid moiety and the 11-and 12-oxo of the alcohol moieties should reduce in the presence of NaBH4 to corresponding hydroxyl groups yielding diol ester structures. Purified Fraction VI, dissolved in isopropanol, was reacted with NaBH, at 40 "C for 30 min, at 60 "C for 60 min, and 82 "C for 120 min. Two reaction products were observed by TLC of the mildly reacted (40 "C) sample (Fig. 8, lane 2 material was the major product (Fig. 8, lanes 3 and 4 ) .
The IR spectrum for the R F 0.35 material showed both ketone and ester carbonyl bands at 1700 and 1740 cm", respectively, and a broad band at 3400 cm-' (Fig. 30), characteristic of a hydroxyl function and similar to the band observed for the Cz8 oxoalcohol (Fig. 3B). The more polar NaBH, reaction product (RF 0.22) showed the carbonyl function for an ester at 1740 cm", but not that for the ketone (1700 cm"), and a hydroxyl function at 3400 cm" (Fig. 3E). These data suggested that the R F 0.35 material was partially reduced and contained both a ketone and hydroxyl group and that the R F 0.22 material was the diol ester.
The compounds with the hydroxyl groups obtained by reduction were derivatized with TBDMSTFA and the resulting derivatives were analyzed by GLC-EI-MS. A typical ion plot of the derivatized R F 0.35 material showed three major peaks (Fig. 5B). Peak 3 gave a mass spectrum consistent for the t-BuMezSi derivatives of 3-hydroxybutanoate esters of 11-and 12-oxooctacosanol (Fig. 9A). The molecular ion at m/z 624 was not detected, however, the M -57 ion (representing loss of the t-butyl group) at m/z 567 was present. The base peak at m/z 161 represented the t-BuMezSi derivative of 3-hydroxybutanoic acid less the 57 mass units of the t-butyl group. The t-BuMezSi group on the penultimate hydroxyl of the acid moiety of the ester was also indicated by the presence of a fragment ion at m/z 159. The presence of the 12-oxo isomer of the alcohol moiety was indicated by the a-cleavage fragment ion at m/z 253 and the @-cleavage fragment ion at m/z 357. A corresponding fragment ion at m/z 343 indicated the presence of a lesser amount of the 11-oxo isomer.
Peak 1 of the R F 0.35 material was identified as the t-BuMepSi derivative of mainly 12-oxooctacosanol (Fig. 9B).
The mass spectrum for his derivative showed an intense ion (100%) for M -57 at m/z 481 and a @-cleavage ion at m/z 271. A lesser amount of the 11-oxo isomer was indicated by m/z 257. Peak 2 gave a mass spectrum consistent with a structure derived from dehydration of the 3-hydroxybutanoate ester (Fig. 9C). This degradation product was characterized by a molecular ion at m/z 492, an RCOOH: fragment ion at m/z 87, and a base peak at m/z 69. Characteristic a-and &cleavage fragmentation at the 12-oxo position supported the identification of this material as mainly 12-oxooctacos-any1 2-butenoate. The ion at m/z 481 is believed to be from tailing of peak 1. Peaks 1 and 2 apparently resulted from partial degradation of the derivatized ester in the flash heater. All three mass spectra support the identification of Fraction VI as being oxoalcohol esters of acetoacetic acid.
GLC-MS analysis of the t-BuMe,Si derivatives of the more polar product (RF 0.22) following NaBH4 reduction of Fraction VI showed three major components (Fig. 5C) and an elution profile similar to that of the RF 0.35 derivatives, but the three components possessed longer retention times. A mass spectrum of peak 3 of the R F 0.22 component confirmed that it was the t-BuMezSi derivative of the dihydroxy ester formed by reduction of the 3-oxobutanoate ester of the Cz8 oxoalcohol (Fig. 1OA) Again, all three mass spectra support the conclusion that Fraction VI was a mixture of oxoalcohol esters of acetoacetic acid.
The tendency for 3-hydroxybutanoate esters to undergo oncolumn hydrolysis and dehydration was confirmed by GLC-MS analysis of the authentic octadecyl 3-hydroxybutanoate t-BuMe2Si derivative (Fig. 11). In addition to the presence of peak 3 for the t-BuMe2Si derivative of the intact ester, the total ion plot showed the resolved peaks 1 and 2, the t-BuMe2Si derivative of the CIR alcohol and the 2-butenoic acid ester of the CIS alcohol, respectively (Fig. 50). The mass spectrum for the t-BuMe2Si derivative of octadecyl3-hydroxybutanoate (Fig. 11A) showed a fragmentation pattern similar to that shown previously for the t-BuMe2Si derivative of 12-oxooctacosanyl3-hydroxybutanoate (Fig. 9A); a base peak a t m/z 161, a less intense fragment ion a t m/z 159 that was indicative of a t-BuMe2Si derivative of a penultimate hydroxyl group, and an M -57 fragment ion at m/z 413. The ion a t m/z 327 was probably a rearrangement ion corresponding to the loss of 57 mass units from the t-BuMe2Si derivative of octadecanol. The mass spectrum of the 2-butenoic acid compound was characterized by a base peak a t m/z 87 indicative of the RCOOH; fragment ion (Fig. 12B). An analogous base peak a t mlz 89 has been reported for n-butanoic acid esters of fatty alcohols (13).

Acid Esters
Earlier studies of the surface lipids from M. sexta had indicated that Fraction IX was the most polar lipid constituent hydrolyzed by alcoholic KOH to yield oxoalcohols (1). Fraction IX was identified as a homologous series of c26, C27, and C, oxoalcohol esters of 3-hydroxybutyric acid. These esters had been previously identified in the RF 0.35 band following TLC separation (Fig. 9) of the NaBH,-reduced oxoalcohol esters of acetoacetic acid (Fraction VI).
Fraction IX showed the same IR spectrum as that of the RF 0.35 band (Fig. 30) and when silylated (TBDMSTFA), gave the same GLC-EI-MS elution profile and retention times ( Fig. 5B) as did the RF 0.35 components. The mass spectra previously shown for the three CZR components in the RF 0.35 band, the intact ester (Fig. lOA), the oxoalcohol hydrolysis product (Fig. lOB), and the dehydrated butenoate derivative (Fig. IOC), were identical to those of the three GLC components derived from Fraction IX (data not shown).
GLC-EI-MS analysis of Fraction IX after treatment with BSA gave mass spectra for the CZR MeaSi derivatives (Fig. 12) with a fragmentation pattern analogous to that observed for the t-BuMe2Si derivative. (It is interesting to note the suppressing effect that the t-butyl group has on the formation and/or intensity of some fragment ions.) The ions at mlz 340 and 130 correspond to rearrangements forming the fragments
However, very little of the butenoate ester of the oxoalcohol was formed. Thus, all the data supported the conclusion that Fraction IX was a mixture of oxoalcohol esters of hydroxybutyric acid.
Fraction VI1 reacted with ethanolic KOH to give a product that migrated on thin layer plates coincident with n-primary alcohol standard and just ahead of Fraction VII. These apparent alcohols were analyzed by GLC-EI-MS as their t-BuMe2Si derivatives and were identified as a series of c26, c27, and C28 n-primary alcohols, with the CZR alcohol as the major component (85%). The mass spectrum for the t-Bu-Me2Si derivative of the CPR alcohol was identical to the mass spectrum shown previously for the t-BuMe2Si derivative of the CP8 n-primary alcohol derived from hydrolysis of total lipid extracts (1).
The ester components in Fraction VI1 reacted with TBDMSTFA and the t-BuMe2Si derivatives were characterized by GLC-EI-MS as c 2 6 " c 2 8 n-primary alcohol esters of 3hydroxybutyric acid. A typical mass spectrum of the t-Bu-MepSi derivative of the CZ8 component showed a molecular ion a t m/z 610, a M -15 at m/z 595, and a more intense ion for M -57 at m/z 553 (Fig. 13). The base peak a t m/z 161 and the fragment ions at m/z 159,203, and 219 are characteristic for the t-BuMezSi derivative of 3-hydroxybutyric acid (see Figs. 9A, lOA, and 1lA). Thus, Fraction VI11 was concluded to be a mixture of the long chain primary alcohol esters of 3-hydroxybutyric acid.

DISCUSSION
The major lipid constituents of the surface wax of diapausing M . sexta pupae were clearly established as long chain oxoaldehydes and corresponding oxoalcohols that existed as alcohol moieties of structures that were susceptible to alkaline hydrolysis (I). The data presented here established the major lipid component (Fraction VI) as oxoalcohol esters of acetoacetic acid. Oxoalcohol esters of acetic acid (Fraction IV) and 3-hydroxybutyric acid (Fraction IX) were also found.
The relative composition of the various lipid components that comprised the surface wax was difficult to determine. Therefore, estimates on the composition of the surface lipids were based on the distribution of radioactivity on TLC plates from lipid extracted from the surface of pupae injected with ['4C]malonate (1). The estimated composition was oxoalcohol esters of acetoacetic acid (35-45%), oxoaldehydes (30-35%), oxoalcohol esters of hydroxybutyric acid (5-lo%), unesterified oxoalcohols (5-lo%), oxoalcohol esters of acetic acid (2-5%), hydrocarbons (2-5%), and unknown fractions including esters of n-primary alcohols (5-10%). The chain length distribution was much the same for oxoaldehydes, oxoalcohols, and nprimary alcohols; CZ6 (&lo%), Cp7 (2-5%), and Cz8 (85-95%). The distribution of oxo isomers, estimated from mass spectral data, varied for the different lipid components. The CB OXoalcohol esters of acetoacetic acid (from Fraction VI) showed a 12-0x0/11-0x0 ratio of 2:1, whereas the ratio for the Czs oxoaldehydes (from Fraction V) was 1:l. The c 2 6 components for all fractions were mainly the 11-oxo isomer and the Cp7 components contained nearly equal amounts of the 11-and 12-oxo compounds. Both the acetoacetate and 3-hydroxybutyrate esters of oxoalcohols were susceptible to thermal degradation. The data  2). Base peaks in all spectra were at mlz 75. The fragment ion peak at mlz 143 in spectrum C had a relative intensity of 83%.
presented in Figs. 5A, 6, and 7 indicate that the acetoacetate esters (Fraction VI) underwent pyrolytic degradation during GLC, GLC-EI-MS, and solid sample probe EI-and CI-MS analyses. In addition to hydrolysis of the t-BuMezSi derivatives of the 3-hydroxybutyrate esters during GLC and GLC-MS analyses, we observed the elimination of the penultimate hydroxyl group from the acyl moiety to form butenoate ester compounds (see Figs. 5B and 9). Since the acetate esters (Fraction IV) were stable to these procedures, the thermal instability of the acetoacetate and 3-hydroxybutyrate esters was apparently due to the presence of the oxo or hydroxyl function at a position p to the ester carbonyl.
Hydrolysis of the total surface lipid extracts followed by GLC-MS analyses had indicated the presence of C2&,, primary alcohol moieties with the same chain length distribution as the oxoalcohols (1). Primary alcohol esters of 3-hydroxybutyric acid were identified in Fraction VII. GLC-MS analyses of fractions recovered from TLC of total lipid revealed the presence of small quantities of acetate esters of primary alcohols in the regions of the TLC plates between hydrocarbons (Fraction I) and acetate esters of oxoalcohols (Fraction IV). Acetoacetate esters of primary alcohols were not identi-fied, however, alkaline hydrolysis of lipids from regions of TLC plates between Fractions V and VI yielded small quantities of primary alcohols. The RF value of these alcohol esters was about the same as that for authentic octadecyl 3-oxobutanoate suggesting that the acyl moieties of the primary alcohols may have been acetoacetate.
To our knowledge, the natural occurrence of fatty alcohol esters of acetoacetic or 3-hydroxybutyric acids has not been reported. However, esters of n-butyric acid were identified as the major volatile constituents of skin secretion of a South American primate, the marmoset, Saguinus fuscicollis (13). The alcohol moieties were identified as a series of CI6-Cz4 saturated and unsaturated fatty alcohols. Acetate esters of fatty alcohols have been found in insect surface lipids. Many of the female sex pheromones of Lepidoptera have been identified as acetate esters of C8-CI6 unsaturated fatty alcohols (14,15). Acetate esters of shorter chain alcohols C,-C,, have been identified as insect exocrine products (16). An acetate ester of a secondary alcohol, 8-heneicosanol, comprised 27% of the surface lipid extracted from the male little house fly, Fannia canicularis (17). cis-Vaccenyl acetate was present in cuticular lipid extracts from male fruit flies, Drosophila melanogaster (18).
Hydrocarbons comprise a majority of the cuticular lipids of many insects, and wax esters, sterol esters, fatty alcohols, and free fatty acids are also common components (19). The cuticular lipids of diapausing M. sexta, however, contain only 4% hydrocarbon (5). The composition of mainly oxoaldehydes and oxoalcohol esters of acetoacetic acid and 3-hydroxybutyric acid represents a unique mixture of polar lipid compounds whose primary function is to prevent desiccation and provide protection from the environment during diapause. Previous studies have indicated that diapausing lepidopteran pupae possess thicker wax layers than nondiapausing pupae (20,21). Scanning electron microscope studies of the epicuticular surfaces of diapausing and nondiapausing pupae of M. sexta have shown that in nondiapausing pupae, the wax layer was smooth, thin, and transparent so as to reveal the underlying cuticulin layer (20). In diapausing pupae, the wax layer was rough, crusty and the underlying cuticulin layer was not visible, indicating a thick wax layer.
Preliminary TLC analysis of the surface lipids extracted from nondiapausing M. sexta pupae showed the same distribution of components as that shown for diapausing pupae (data not shown). Therefore, the main difference in the wax layer of diapausing and nondiapausing pupae is not in the quality of the wax but in the quantities of wax secreted on the epicuticle. The physiological state of diapause in insects is under hormonal control (22,23) and the deposition of additional wax may result from hormonal changes that accompany diapause induction. Elucidation of the biosynthetic pathways involved in the formation of these novel surface lipid components may provide for a better understanding of possible hormonal regulation of the synthesis and deposition of surface wax. Furthermore, the utilization of increased quantities of acetoacetic and 3-hydroxybutyric acids as acyl moieties of major components of cuticular lipids may be of metabolic significance to the insect during diapause.