Lipids of the tail gland, body and muzzle fur of the red fox, Vulpes vulpes

Electronic Supplementary Material

Stuart McLean, Noel W. Davies, David S. Nichols

Analyses of mono-alcohols, glyceryl ethers and fatty acids

The TMS ethers of alcohols and 1,2-diols of the same carbon number produced an ion of the same mass and abundance (100%), corresponding to [M – 15]+ for the alcohol and [M – 103]+ for the 1,2-diol. However, the alcohols lacked ions at m/z 147 and 205 and, unlike the 1,2-diols, the m/z 75 ion was greater than m/z 73. Additionally, the alcohol TMS eluted considerably (about 180 Kovats units) before the corresponding n-1,2-diol TMS2. The alcohols found were n-isomers of C24, 26-28 and C30 (Table 2).

Although the amounts of FA were greater in the acid extracts, some were found in the neutral extracts, where they interfered with the MS of several 1,2-diols. For example, a symmetrical although broad TIC peak at KI 2254 showed ions m/z 117 and 147 as well as at m/z 313, 341 and 356 which were not interpretable as a single compound. Extracted ion chromatograms partially resolved m/z 313 and 341, and comparison of the MS of the first and second half, respectively, of each peak showed the presence of a 17:0 1,2-diol TMS and an 18:0 FA TMS. Diol ions were m/z 313 [M – 103]+ (100%), 410 [M – 15]+ (2.7%), 147 (19%) > 145 (2.5%), and the FA ions were m/z 341 [M – 15]+ (100%), 256 M+ (19%), and 117 (88%). In each case, the contaminating ions from the other compound were less than in the other MS. Since both 1,2-diols and FA showed the same isomeric pattern of iso-, anteiso- and n-, this coelution was a common occurrence.

The GE (TMS) showed the ions characteristic of other 1,2-diols (TMS): m/z 205 (base peak in this case), and abundant ions at m/z 73, 103, 117, 133 and 147. The mass of each GE was obtained from [M – 15]+ (loss of CH3), [M – 90]+ (loss of OTMS + H), and [M – 147]+ (loss of TMS2 + H). KI values were compared with those found previously in our laboratory [10, 20]. For example, 1-O-octadecylglycerol bis-TMS (the most abundant GE) produced ions at m/z 205 (100%), 73 (35%), 103 (16%), 117 (38%), 133 (33%) and 147 (39%), as well as 473 (4.1%), 398 (3.7%) and 341 (16%) from losses of 15, 90 and 147 daltons respectively. KI was 2688 (previously 2669). GE were found with O-alkyl groups C12 and C15-19, mostly n- (Fig. 2, ESM) but with some branched isomers and one unsaturated compound (18:1).

Analyses of fatty acids as methyl esters (FAME)

This section gives details on FAME identifications not covered in the main article.

Iso- and anteiso-isomers

The EI MS confirmed the iso-isomers which showed the expected [M − 43 (C3H7)]+ ion. which was greater than [M − 29 (C2H5)]+ or [M− 57 (C4H9)]+. However, the reverse was not always the case for anteiso-isomers where [M − 43 (C3H7)]+ was frequently more abundant. Although this is not an obvious loss considering the ω-2 methyl branching of anteiso-isomers, this loss can also occur from C2-4 by a complex rearrangement at the ester end of the molecule [13]. Here the collision-induced dissociation (CID) data showed their value.

2-Methyl isomers

The MS of 2-methyl branched FAME were very similar to those of the ethyl esters of the Cn-1 FA, but the latter lack the [M − 31]+ ion and elute about 40 KI units later than the 2-methyl FAME.

4-Methyl isomers

The series KI 1534-2134 eluted 37 KI units earlier than the corresponding 4-monomethyl isomers, suggesting another branch at the iso-position. This was supported by the CID MS which showed a more abundant [M − 43]+ ion in this series than in the 4-monomethyl series. For example, 17, 4-, i- (KI 1934) showed a more abundant [M − 43]+ ion at m/z 241 (11% ) than did 17, 4- (KI 1971) (2%). This ion was not significant in the EI MS of either compound, showing the value of using the more selective CID mode for analysis of branching positions beyond C4.

Similarly, the remaining series (KI 1641-2043), offset by 30 KI units from the 4-monomethyl isomers, was identified as 4-methyl-anteiso-FAME. For example, 18, 4-, ai- (KI 2043) showed a more abundant [M − 29]+ ion at m/z 269 (23%) than did 18, 4- (KI 2071) (14%). Again, the EI MS gave weaker m/z 269 ions in both compounds (4.1 and 1.4%, respectively).

Other branched methyl groups

Branching around the middle of the FAME carbon chain was explored by searching for the two expected COM fragmentation ions, 28 daltons apart, formed either side of the methyl-branch carbon [13, 14].The ion formed from the ester end of the molecule to the branch [a]+is accompanied by [a + 1] + and [a + 2] + ions. The ion formed by cleavage after the branch [b]+ also has confirmatory ions: [b − 32] + and [b – 32 − 18] +. Searches were made for candidate [a] + and [b] + ions, either side of a very diminished or absent intermediate COM ion, but in many cases the pattern was equivocal, due either to the low abundance of [a] + and/or [b] +, the significant abundance of the intermediate ion, or interferences from co-eluting FAME. Many of the mid-chain branched isomers elute closely together [14, 17], which made their KI values less useful and resulted in peaks with mixed mass spectra which were less interpretable.

A series of 6-methyl FAME (KI 1665-2063) was, however, identified by their prominent [M – 76]+ and [M – 75]+ ions, thought to be due to a loss of a hydrogen at C6 and subsequent rearrangements leading to losses of CH3OH (32) and C2H4O (44) [13]. For example, the EI MS of FAME 15, 6- (KI 1764) gave m/z 180 [M − 76]+ (69%) and m/z 181 (10%), as well as other ions indicative of the C6 cleavage at m/z 206 [M − 50]+ (5%), 115 [a] + (19%), 116 [a + 1] + (5%), 143 [b] + (13%), 111 [b − 32]+ (14%) and 93 [b − 50] + (2%). The CID data confirmed the [a]+ and [b]+ ions and weak [a + 14]+ ion (Table 4). FAME 17, 6- (KI 1963) eluted in a complex split peak in which the 8- and 10-methyl isomers eluted together in the second part (KI 1966). Their combined MS showed [a]+, [a+14]+ and [b]+ ions for both 8- (m/z 143, 157, 171) and 10-methyl isomers (m/z 171, 185, 199), in both EI and CID (Table 4). In addition, the EI MS had a doublet of m/z 129 (15%) and 130 (10%) which is characteristic of 10-methyl isomers [14]. The calculated KI values for a methyl branch at positions 6-, 8- and 10- in FAME 17 were within one KI unit (Table 4), showing that co-elution is inevitable.

Two other FAME (KI 1726 and 1826) showed characteristic 6-methyl-branching ions, but eluted 38 KI units before the corresponding 6-methyl isomers, indicating that they were 6-methyl-iso-isomers. The CID mass spectrum of 15, 6, iso- showed a strong [M − 43] + ion at m/z 213 (44%), compared with 0% in 15, 6-, supporting the iso-structure.

FAME KI 1879 and 1973 showed clear evidence of a 12-methyl branch in both the EI and CID mass spectra (Table 4). Two other FAME, KI 1829 and 2030, which eluted 50 KI units earlier, showed similar evidence of a 12-methyl branch, but their expected additional branch could not be determined from their mass spectra. MS of FAME KI 2081 and 2174 showed the ions which indicate a 14-methyl branch, more clearly in their CID than EI spectra (Table 4).

Scheme 1 Major fragmentation pathways in the mass spectrum of 2-hydroxy-nonadecyl-iso-valerate. The mass spectrum is shown in Fig. 3A of the main article

Scheme 2. Major fragmentation pathways in the mass spectrum of 2-hexadecanoyloxy-octadecyl-iso-valerate. The mass spectrum is shown in Fig. 5 of the main article.

1