Introduction

Most waterfowl that forage underwater do not have to dry their feathers after diving or dowsing due to the water-repellent nature of their plumage. This water repellency is partly due to the water-repellent properties of keratin but also to the geometrical structure of the feathers (Rijke 1968; Bormashenko et al. 2007; Rijke and Jesser 2011; Srinivasan et al. 2014). In particular, the distance between the barbs and barbules seems to be very important: the smaller the distance, the more difficult it is for water to penetrate the surface of the vane (Rijke 1970). In addition, the secretion of the uropygial gland that is spread onto the feathers during preening has been implicated in the water repellency of feathers (Jacob and Ziswiler 1982; Spearman and Hardy 1985; Chiale and Montalti 2013). Furthermore, the uropygial secretion seems to play an important role in olfactory communication, signaling, and crypsis, in feather integrity and in defense against fungi, bacteria, and parasites, and its chemical composition varies seasonally (Kolattukudy et al. 1985, 1987, 1991, Grieves et al. 2022); for further references see Stangier et al. (2023).

In contrast, the Great Cormorant, Phalacrocorax carbo (Phalacrocoracidae), shows a typical wing-spreading behavior, presumably to dry its feathers (Johnsgard 1993; Grémillet et al. 2005; White et al. 2008), because the feathers are partly wettable (Rijke 1968; Rijke and Jesser 2011; Srinivasan et al. 2014). This wettability causes lower buoyancy, thus increasing diving performance (Grémillet et al. 2005; Ribak et al. 2005). It has been linked to the unusual structure of the Cormorant’s contour feathers: close to the rachis, the vane is closed whereas in the periphery, the vane is open because the barbules are not interconnected (Rijke 1968; Rijke and Jesser 2011; Srinivasan et al. 2014; Stangier et al. 2023). The closed vane provides an insulating air layer whereas the peripheral open vane causes the partial wettability (Grémillet et al. 2005). Nevertheless, a contribution of the uropygial secretion to the Cormorant’s water repellency has not been ruled out.

The uropygial gland is especially well developed in aquatic birds and contains a water-repellent secretion comprising a variety of carboxylic acid or wax-type esters (Jacob 1978; Salibian and Montalti 2009), but also other components, such as alcohols, ketones or immune proteins (Jacob et al. 1997; Whittaker et al. 2010; Amo et al. 2012; Carneiro et al. 2020). The first report about the uropygial secretion’s chemical constituents established carboxylic acid esters as major components, exemplified by octadecyl palmitate and stearate (Stern 1905). Since then, the uropygial gland secretion has been investigated in various species using modern coupled mass spectrometric techniques, usually gas chromatography/mass spectrometry (GC/MS). A great variety of wax esters of medium- to long-chain fatty acid esters with and without methyl branches along the chain have been identified in uropygial secretions (Jacob 1978; Priol et al. 1991; Hoerschelmann and Jacob 1992; Salibian and Montalti 2009; Rajchard 2010).

We describe here the composition of the lipophilic part of the secretion of the uropygial gland of the Great Cormorant, applying a new mass spectrometric technique, atmospheric solids analysis probe-atmospheric pressure chemical ionization-mass spectrometry (ASAP-APCI-MS) on bird-wax analysis. The new method, ASAP-APCI-MS (McEwen et al. 2005), allows a fast and easy analysis of the secretion without alteration of the sample on a relatively inexpensive mass spectrometer. The sample is directly introduced via a glass tip and all mass spectrometric ions of the secretion up to a molecular mass of about 1500 Da are recorded, resulting in a mass spectrum of the whole secretion. In contrast, conventional gas chromatography/mass spectrometry (GC/MS) includes chromatographic separation that reveals mass spectra of individual components of the secretion, but it has an upper range limit of about 800 Da and a tendency to discriminate against compounds of higher molecular weight. Therefore, ASAP-APCI-MS gives an overview of the overall composition of the secretion according to molecular mass, while GC/MS gives detailed structural information of the compounds in the secretion, especially when combined with derivatization. We applied both mass spectrometric methods here to get a clearer picture of the complex composition of the secretion. The lipophilic part of the Great Cormorant’s uropygial gland contents has, to the best of our knowledge, not been analyzed before. The physical structure of the feathers is discussed in the accompanying contribution (Stangier et al. 2023) and earlier reports (Rijke 1968; Grémillet et al. 2005; Srinivasan et al. 2014).

Although many reports on bird waxes have been published, their results are not in all cases comparable to our results because of slight differences in the protocol used, sensitivity, accuracy or instrumentation. Therefore, we decided to include the secretion of the Muscovy Duck, Cairina moschata (Anatidae), in our study as a reference. Their wax composition and that of other ducks has been intensively investigated (Odham 1967; Jacob et al. 1979; Bertelsen and Nguyen 1982; Livezey et al. 1986; Yu et al. 1989; Jakob and Hoerschelmann 1993). We here describe a striking difference between the two species, not only in wax composition, but also in the number of found compounds. Water repellency of the feathers is not only influenced by the waxy layer originating from uropygial secretions, but also by the surface structure of the feathers. These two elements are difficult to disentangle, but for a better understanding of this cooperative effect, it seemed to be important to measure the water repellency efficiency of feathers and secretions. A straightforward approach is contact angle measurements of water, which is an indicator for water repellency. A correlation with the chemical composition of preen oil would also allow insight into structural or compositional effects of compounds or mixtures especially suited for secretion and for the feathers to get an insight into the relative contributions of the two factors to the observed repellency. Therefore, we performed contact angle measurements of water with the secretions of the uropygial gland as well as with feathers before and after the removal of the natural wax layer to investigate their water-repellent properties.

Methods

General procedures

All chemicals used were bought in reagent grade purity from Sigma-Aldrich, Roth, and Acros and used without further purification if not stated otherwise. All solvents used were of HPLC grade or of higher quality.

Animals

The uropygial glands of three adult Great Cormorants (two males, one undetermined) were used for the biochemical analysis of the uropygial secretion. The animals were shot in the course of a lethal control measure during the first two weeks of February 2016 (permit 67.1-2.03.20-pi, 09-2014), Untere Naturschutzbehörde Rhein-Sieg-Kreis, and permit 35/61.92.16 (001/16), Untere Landschaftsbehörde Hochsauerlandkreis), i.e., before the breeding season starting in March/April. One animal was dissected within 24 h post-mortem, the other two birds were deep frozen until use. The glands were dissected, measured, weighed, and fixed in acetone. The weight of the intact glands was 3.0, 3.4, and 3.6 g, respectively. For comparison, the uropygial gland of a male adult Muscovy Duck obtained from a commercial breeder in March 2016 was dissected within 1 h post-mortem. The gland weighed 5.7 g.

Sample preparation

The secretion was extracted from the glands using a 150 mL Soxhlet extractor with chloroform/methanol (2:1 v/v) as extraction solvents at reflux conditions (Dobush et al. 1985). In the Cormorants, the acetone extraction yielded 0.401, 0.378, and 0.423 g, the chloroform/methanol extraction yielded 1.139, 1.053, and 1.204 g extract, respectively. For the Muscovy Duck, we obtained 0.559 g extract from the acetone and 1.2 g extract from the chloroform/methanol extraction. Both extracts of each sample, from acetone and chloroform/methanol, were subjected to thin layer chromatography (TLC) to pre-determine their general composition. TLC was performed on silica with light petroleum/diethyl ether/glacial acetic acid (90/10/1 v/v) and the samples stained with cerium sulfate staining solution (2% w/w Ce(SO4)2/16% v/v H2SO4) (Jacob and Grimmer 1970). The two extracts were combined due to a generally comparable composition of waxes, triglycerides, and sterols. Solvents were removed with a rotary vacuum evaporator (Büchi Rotavapor RE-111). The secretion was stored in two portions per gland in centrifuge tubes at −20 °C until use. For analysis, a small portion, about 5 mg of the secretion, was dissolved in 200 µl dichloromethane (DCM) and filtered through a pipet plugged with laboratory tissue prior to derivatization or injection into the gas chromatograph. For direct injection-electrospray ionization-high resolution mass spectrometry (DI-ESI-HRMS) analysis, the sample was dissolved in DCM/methanol (1:1) and filtered as stated above.

Direct atmospheric solids analysis probe-atmospheric pressure chemical ionization-mass spectrometry (ASAP-APCI-MS)

A direct probe, a small glass stick, was introduced into the wax sample and then directly into the ASAP-ionization chamber of an Expression MS (Advion) mass spectrometer equipped with an APCI ionization unit operated in positive ionization mode. The capillary temperature was 250 °C, the capillary voltage was 180 V, the source voltage offset was 20 V, and the source voltage span was 30 V. APCI source gas temperature was 350 °C and the APCI corona discharge was set to 5.0 µA. The scan range was 118–1522 amu.

Gas chromatography/mass spectrometry (GC/MS)

GC/MS analyses were carried out on a GC 7890A gas chromatograph connected to a 5975C mass-selective detector (Agilent Technologies, USA) with an HP-5ms fused silica capillary column (30 m, 0.22 mm i.d., 0.25 μm film, Agilent Technologies, USA). Conditions were as follows: carrier gas (He): 1.2 ml/min; injection volume: 1 µl; injector: 250 °C; transfer line: 300 °C, EI 70 eV. The gas chromatograph operated in splitless mode and the temperature program was as follows: 50 °C (5 min isothermal), increasing with 5 °C/min to 320 °C, which was held isothermal for 5 min. The scan range was 34–700 amu. Linear retention indices were determined from a linear, homologous series of n-alkanes.

Compound identification

Compounds were identified using mass spectrometry and gas chromatography. Individual compounds were identified by interpretation of EI-mass spectra according to established rules. Linear gas chromatographic retention indices were measured and compared to theoretical values obtained by an empirical increment system described earlier (Schulz 2001). Extracts were analyzed directly by GC/MS or derivatized into methyl esters, pyridin-3-ylmethyl esters, or nicotinic acid esters before analysis. A more detailed description of the analytical procedure can be found in the results section below.

Direct injection-electrospray ionization-high resolution mass spectrometry (DI-ESI-HRMS)

Measurements were performed using a Velos LTQ Orbitrap (Thermo Scientific) mass spectrometer with a custom-made direct injection ESI microspray device. Flow rate was 1 µl/min through a stainless-steel capillary (90 µm i.d.) in positive ionization mode. Scan range was set to 100–2000 amu and spray voltage was 2.09 kV.

Hydrolysis of fatty acid esters

A fraction of the solid extract (1 mg) was dissolved in 200 µl KOH/methanol (10%) in a 2 ml vial, equipped with a stirring bar. This mixture was stirred at 40 °C for 24 h. Workup was carried out by adding 2 M HCl until the pH of the reaction reached 2, followed by three times extraction with DCM. The combined organic phases were dried over anhydrous Na2SO4 and filtered. Hydrolyzed esters were redissolved in 1 ml DCM.

Synthesis of methyl esters

Trimethylsulfonium hydroxide solution (TMSH) (50 µl, 0.25 M in MeOH) was added to a solution of 0.2 mg hydrolyzed esters in 200 µl DCM in a 2 ml vial. After 1 h at room temperature (rt), the solvent and leftover TMSH was removed under a stream of N2 and the sample was redissolved in 100 µl DCM (Müller et al. 1990).

Synthesis of nicotinic acid esters

A solution of 0.2 mg hydrolyzed esters in 200 µl DCM were treated with 1 mg nicotinic acid, a catalytic amount of 4-(N,N-dimethylamino)pyridine (DMAP) and 0.5 mg 1-ethyl-3-[3-(dimethylamino)-propyl)]carbodiimide hydrochloride (EDC) in a 2 ml vial. The reaction mixture was stirred at rt for 24 h. Then the solution was washed twice with 2 M HCl, followed by a wash with saturated NaHCO3 solution. The organic phase was dried over anhydrous NaSO4 and filtered prior to GC/MS analysis.

Synthesis of pyridin-3-ylmethyl esters

A slightly improved procedure of Schulze et al. (2017) was used for esterification with 3-pyridinemethanol. A solution of 0.2 mg hydrolyzed esters in 200 µl DCM was treated with 20 µl of oxalyl chloride in a 2 ml vial. After 1 day, excess oxalyl chloride and the solvent were removed and the sample was redissolved in 100 µl DCM. For esterification, catalytic amounts of DMAP and one drop of freshly distilled 3-pyridinemethanol were added. The reaction mixture was heated to 60 °C for 1 h and filtered. The filtrate was washed with H2O three times, the organic phase was dried over anhydrous Na2SO4 and filtered prior to GC/MS analysis.

Feathers

Feathers of Muscovy Ducks were opportunistically collected from colonies kept in Zoo Karlsruhe and Gruga Essen. Cormorant feathers came from the Cormorant carcasses used in this study (see also Stangier et al. 2023).

Contact angle measurement

Static contact angles were measured with a DATAPHYSICS Type OCA 15 optical tensiometer. Millipore water (9 µl) was used as liquid phase at rt. The liquid phase was applied to the substrate surface using the sessile drop method and the static contact angles were measured using a camera and the SCA 20 software. Two to three replicates of each sample were measured two to three times. Substrates were dried in a stream of N2 between measurements.

Two portions of the gland extract of each species were dissolved in DCM to prepare solutions with concentrations of 25 and 50 mg/ml for each species. Each sample was applied to a titanium surface by spin coating using 60 µl of sample solution in a custom-made device at 2000 rpm for 30 s. The titanium surfaces were cleaned by sonication in DCM prior to use. Blanks were prepared using only DCM.

Contour and flight feathers of both species were cut in pieces if needed to get even surfaces. Three even spots on three feathers were chosen for the measurement. To remove the lipidic coating, feathers were placed in a beaker with pentane for 2 h to extract feather waxes and were carefully rinsed with pentane and air dried afterward. Contact angles of feathers were measured before and after washing with pentane.

Measured contact angle values were cleaned for outliers. For this purpose, an outlier test according to Dean and Dixon was performed (Rorabacher 1991). Then an f test and a t test were carried out with the cleaned data. Both tests were performed with the help of the Analysis ToolPak Excel Addin. All tests were carried out using a p value of 5%.

Results

Dichloromethane extracts of uropygial gland contents obtained from Great Cormorants and for comparison, a Muscovy Duck, were investigated by mass spectrometry. While GC/MS of natural gland secretions and its derivatization products is the method of choice for the analysis of bird gland waxes (see e.g., Jakob and Hoerschelmann 1993; Dekker et al. 2000), we additionally introduced ASAP-APCI-MS (McEwen et al. 2005) that allows to get a quick overview of the contents of a sample up to 1500 Da, extending the limitation of analyte size in GC/MS, which is usually around 700–800 Da. Therefore, additional secretion components might be detectable by ASAP-APCI-MS that evade GC/MS analysis, e.g., common fats such as triglycerides that have a molecular mass between 800 and 900 Da.

ASAP-APCI-MS analysis

ASAP-APCI-MS analysis in positive ionization mode showed different sets of compounds for Duck and Cormorant. In the low-molecular-weight region, peaks dominate that originate from the fragmentation of wax esters or alcohols due to elimination of an acid or water. In the Cormorant’s extracts, predominantly [M + H]+-ions were observed originating from esters ranging in molecular weight from 400 to 580 Da, as well as a second set of compounds in the range of 900–1100 Da, which are likely dimeric esters [2M + H]+ (Fig. 1A), a mass spectrometric artefact. The molecular ions are more or less evenly distributed within these mass ranges. A small amount of compounds between 800 and 900 Da occurs, matching the molecular mass of common triglycerides. In the extract of the Muscovy Duck, esters with a molecular weight of 400–500 Da were detected, and, in addition, a second set of compounds in the range 550–620 Da (Fig. 1B).

Fig. 1
figure 1

Mass spectrum of A Great Cormorant and B Muscovy Duck uropygial gland contents obtained by ASAP-APCI-MS in positive ionization mode. Cleavage products are ions below m/z 300, likely acids and alkenes that are formed by fragmentation of the esters

Full size image

In comparison, the Cormorant seems to produce a broader range of compounds than the Muscovy Duck. The esters from the Cormorant have a higher molecular weight and a much more homogenous distribution. Fast analysis of all our samples showed high similarity within the three Cormorant samples. ASAP-APCI-MS seems to be a promising tool for the initial analyses of complex wax mixtures, and will have its strength in the fast comparative analysis of a large number of samples. This method immediately shows molecular weight distributions within the sample. In contrast, it offers very limited possibilities for structure elucidation, which require moderately expensive GC/MS systems or high-end instruments useful for high-resolution mass spectrometry as discussed now. The ester profile of the Cormorant contained many more individual compounds than that of the Muscovy Duck. Compounds in the higher mass range detected in both secretions were not detected by GC/MS, likely due to their low volatility.

High-resolution mass spectrometry

By use of direct injection-electrospray ionization-high resolution mass spectrometry (DI-ESI-HRMS), the molecular composition of the Cormorant’s compounds was determined to predominantly include compounds of the sum formula CnH2nO2Na+, typical for saturated carboxylic esters and acids, with n ranging from 25 to 45 (supplementary Fig. S7). After hydrolysis of the sample, the observed peaks disappeared and lower molecular weight peaks occurred. This indicated that the compounds were esters, which were cleaved into their alcohol and acid part.

Muscovy Duck extracts were dominated by two compounds of m/z 461.43301 and m/z 489.46414 with the molecular formulas C29H58O2Na+ and C31H62O2Na+ with an error of 0.234 and −0.127 ppm (supplementary Fig. S8). The obtained mass spectra of the secretions confirmed the results obtained by ASAP-APCI-MS and were in line with our GC/MS results that are discussed in the next section.

GC/MS analysis

The GC/MS analysis of the extracts revealed large differences between the two bird secretions (Fig. 2). While in the Muscovy Duck, some specific compounds are dominant, in Cormorants a large number of compounds with similar masses were observed, resulting in a chromatogram that showed a broad, not well-structured peak pattern, also often found in heavily oil-contaminated analytical samples. Nevertheless, similar chromatograms were reported from Sanderling, Calidris alba (Dekker et al. 2000), and proved to be a complex mixture of hundreds of compounds. Detailed analysis of the mass spectra showed that the Cormorant compounds were indeed not contaminants, but predominantly esters of fatty acids as expected. They were identified by their typical mass spectrometric fragments, CnH2n+ (alkene from alcohol part), CnH2n+1O2+ (acid + H), and CnH2n-1O+ (acylium ion) (supplementary Fig. S9) (McLafferty and Turecek 1993; Francke et al. 2000; Chinta et al. 2016).

Fig. 2
figure 2

Total ion chromatograms (TICs) of the GC/MS analysis of Great Cormorant (A) and Muscovy Duck (B) uropygial gland extracts. The peak at 19 min is ethyl 2-phenylpropanoate

Full size image

Previous analyses of complex uropygial gland secretions have been performed using extensive GC/MS and GC/MS/MS methodology, e.g., on Calidris sp., allowing 72 different esters to be assigned (Dekker et al. 2000). Nevertheless, the Cormorant extracts are of even higher diversity. This can be seen by the extracted ion traces shown in Fig. 3. The characteristic [acid + 1]+ ions (CnH2n+1O2+, see supplementary Fig. S9) in EI-mass spectra can only be formed by a single fragmentation pathway (McLafferty and Turecek 1993). Therefore, these ions can be used as indicator for characterizing the acid parts in the wax esters with an identical number of carbon atoms. The respective ions for C12 and C14 acids, m/z 201 and 229, show a wide distribution along the time axis with more than 150 peaks, albeit sometimes of low intensity. Isomeric esters of identical mass are sometimes not resolved in GC. Given that this co-elution occurs, one can roughly estimate that there are more than 200 compounds with the same sum formula of the acid part. The same is true when looking at the molecular ions, exemplified in Fig. 3 by m/z 424 (C28) and 452 (C30). The distribution here is narrower, because of the gas chromatographic elution on an apolar GC phase that separates roughly according to compound size for similar compounds. We conclude that calculated conservatively certainly more than 1000 individual wax esters constitute the Cormorant’s gland secretion. This complex bouquet also prohibits identification of individual wax ester species in the natural samples by simple GC/MS, as has been shown earlier for similarly complex samples. As an example, Fig. 4 shows the mass spectrum of a single scan of the Cormorant secretion. The molecular ions (red) indicate at least four different compounds in this peak with different molecular weight, while the acid parts (blue) show at least five. This makes proper identification difficult and usually requires better resolution, e.g., by GC/MS/MS (Dekker et al. 2000). In addition to the wax esters, the Cormorant samples contained one more volatile compound that was identified to be ethyl 2-phenylpropanoate (Fig. 2A at 19 min).

Fig. 3
figure 3

TIC and extracted ion chromatograms (m/z values) of an extract of the uropygial gland of the Great Cormorant. Each peak in each trace indicates at least one ester, sometimes even more due to overlapping retention times of different esters

Full size image
Fig. 4
figure 4

Mass spectrum of scan 6133 at retention time 49.105 min of the GC/MS chromatogram of the Great Cormorant (Fig. 2A). The red ions indicate the molecular mass of the analytes, showing the presence of at least four different esters with different molecular formulas. The characteristic blue [acid + 1]+ ions indicate the presence of at least five different acid parts within these esters

Full size image

By contrast, in the Muscovy Duck, two major peaks are detectable in the TIC that were identified as octadecyl and eicosyl 2,4,6-trimethyloctanoate (Fig. 5, peaks 1 and 2 in Fig. 2B). These compounds are the major lipids, while all other compounds detected are present in much lower quantity.

Fig. 5
figure 5

EI-mass spectra of the two main compounds in the extract of Muscovy Duck. A Spectrum of 2,4,6-trimethyloctanoic acid octadecyl ester (Peak 1 in Fig. 2B); B spectrum of 2,4,6-trimethyloctanoic acid eicosyl ester (Peak 2 in Fig. 2B)

Full size image

Derivatization of extracts

With the methods described so far, a general overview of the composition of the extracts was obtained, but a more detailed structural analysis required cleavage of the esters into their constituting alcohols and acids to allow individual compound identification. Therefore, wax esters (A in Fig. 6) were cleaved into acids (B in Fig. 6) and alcohols (C in Fig. 6), which were subsequently esterified with trimethylsulfonium hydroxide (TMSH) (Müller et al. 1990) into shorter methyl esters (D in Fig. 6). Bird waxes often contain methyl-branched esters (Jacob 1978; Kolattukudy et al. 1991) that are easier to structurally characterize by converting them into the well-known methyl esters. Furthermore, the transesterification process reduces the number of compounds, making the analysis less complex. Localization of methyl substituents in the methyl ester works especially well if the methyl group is located near the ester group because the intensity of characteristic fragments resulting from cleavages next to the methyl group is increased (Ryhage and Stenhagen 1960; Schulze et al. 2017). We were able to identify several methyl-branched compounds, mostly with methyl groups at C-2 and C-3, e.g., methyl 3,7-dimethyldodecanoate. In its mass spectrum (Fig. S10A), an ion characteristic for unbranched methyl esters, m/z 87 of the ion series [CH3OCOCnC2n]+, is shifted to m/z 101. Together with the unaltered ion m/z 74, arising from McLafferty rearrangement, this localizes a methyl group at C-3. The ion m/z 157 is missing and the adjacent ions have slightly increased intensity, indicating a second methyl branch at C-7. All other fragments arising from chain cleavage have very small intensities.

Fig. 6
figure 6

Derivatization reactions performed for the structural characterization of naturally occurring wax esters. Wax esters (A) were hydrolyzed into acids (B) and alcohols (C) and subsequently derivatized into methyl esters (D), pyridin-3-ylmethyl esters (E), and nicotinic acid ester (F)

Full size image

Thus, analyzing the obtained methyl esters of the secretions, 75 different acids were found in the Cormorant and 48 acids in the Duck samples (Table 1), together with 48 alcohols in the Cormorant and 17 in the Duck (Table S1 and Table S2). The carbon chain length for acids and alcohols in the Cormorant includes a significantly greater length variance than in the Duck. Surprisingly, acids with aldehyde and keto groups as well as five dicarboxylic acids, ranging from C-7 to C-11, were detected in the Duck secretion. With the duck’s ability to form diesters, we suggest that the compounds detected in the 550–620 Da range in ASAP-APCI-MS might be diesters of these dicarboxylic acids (Fig. 1B). Unfortunately, methyl groups in the middle and at the end of the alkyl chains could not be unambiguously identified from the mass spectra of the methyl esters or alcohols. The biggest problem, especially in compounds occurring at low concentrations, is the low fragment intensity of characteristic ions and multiple possible fragmentation pathways. Furthermore, methyl esters with more than one methyl substituent were observed that elude identification based on their mass spectra and retention index.

Table 1 Summary of identified alcohols and acids in the Muscovy Duck and the Great Cormorant
Full size table

We, therefore, used additional derivatization methods leading to pyridylmethyl esters (E in Fig. 6) for the acid part and to nicotinic acid esters (F in Fig. 6) for the alcohol part of the hydrolyzed fatty acid esters. These compounds show an increased stability of indicative ions making location of methyl groups along the chain possible. Pyridin-3-ylmethyl esters were obtained by the reaction of the hydrolyzed acid moiety with 3-pyridinemethanol (Harvey 1982, 1984; Chinta et al. 2016), while nicotinates were obtained by reaction of the alcohol with nicotinic acid (Vetter and Meister 1981; Chinta et al. 2016) (Fig. 6). The ions containing a pyridine ring are more stable under electron impact (EI) conditions, allowing fragmentation only from one side of the compound and restricting it to N-containing ions. Characteristic ions for pyridin-3-ylmethyl esters are m/z 151 and 164. The position of methyl groups along the chain can be identified by a missing peak in the ion series [C6H7NOCOCnC2n]+. As an example, the spectrum of pyridin-3-ylmethyl ester of 3,7-dimethyldodecanoic acid is shown in Fig. S10B. All fragments arising from chain degradation show an increased intensity compared to the spectra of the methyl ester (Fig. S10A). Fragments formed by cleavage next to a methyl group have even higher intensity and peaks of m/z 164 and m/z 234 are noticeably reduced within the mentioned ion series. Additionally, double-bond location is facilitated. Figure S11 shows the mass spectrum of the pyridine-3-ylmethyl ester of 9-hexadecenoate. The location of the double-bond was indicated by the allylic cleavage ions m/z 220 and m/z 274 and a 26 amu gap between ions m/z 234, m/z 260 (Harvey 1982; Christie et al. 1987).

Analysis of the alcohol part of the natural waxes is similar, but based on the use of nicotinic acid esters. Typical fragments for such esters are m/z 107 formed by α-cleavage and the McLafferty-ion m/z 124. The spectrum of 14-methyloctadecyl nicotinate shows both typical fragments, and the methyl group was identified to be located at C-14 due to the 28 amu gap between ions m/z 304 and 332 (Fig. S12).

With these derivatization methods, we were able to identify 34 acids in the Cormorant and 19 in the Duck that we could not identify as methyl esters (Table 1). Methyl substituents in the Cormorant acids were located preferentially at C-2, C-3, and C-4 for monomethyl-branched acids. Double-branched acids carried methyl groups at C-3 together with C-7, C-9, or C-11 as well as at C-4 together with C-8 or C-10 (Table S1). The Muscovy Duck esters contained mostly monomethyl-branched acids, located at C-2, C-4, C-6, C-10, C-12, or C-14. Additionally, a few dimethyl- and trimethyl-branched acids such as 2,4,6-trimethylnonanoic acid were found. Furthermore, 17 alcohols were identified as components of the Duck esters, comprising unbranched alcohols as well as monomethyl acids carrying the methyl group at even-numbered positions between C-10 and C-18. The Cormorant esters contained 48 alcohols with up to 3 methyl groups. Branching at odd-numbered positions between C-3 and C-15 occurred only in fatty acids and alcohols from Great Cormorant (Table S2).

Contact angle measurements

We next explored whether the distinct difference in chemical composition of the lipids, a mixture of more than 1000 individual components from Great Cormorant against a few major distinct lipids from Muscovy Duck, both of similar molecular weight, have an influence on water repellency which is particularly important for waterfowl. Therefore, contact angle measurements of water were performed. In case of a hydrophilic surface, the interaction between the water drop and the surface will be strong, resulting in a reduced contact angle.

Hydrophobic surfaces lead to higher contact angles and less interaction. The average contact angle for Cormorant lipids was 24.7° when using solutions with a concentration of 50 mg/ml and 31.4° for 25 mg/ml on a titanium surface (Table S3). These angles are significantly lower than the Ducks’ contact angles with 52.8° for 50 mg/ml and 48.8° for 25 mg/ml (Fig. 7A, supplementary Fig. S13), indicating a reduced hydrophobicity of the Cormorant’s extracts.

Fig. 7
figure 7

A Contact angles for Muscovy Duck and Great Cormorant uropygial gland extracts at concentrations of 25 and 50 mg/ml on a titanium surface. B Contact angles for Muscovy Duck and Great Cormorant feathers before and after treatment with pentane. Bars with different labels have significantly different means at the alpha = 5% level

Full size image

Additionally, we measured the contact angle of the birds’ feathers (Fig. 7B). The average contact angle of Muscovy Duck feathers was 136.3° and that of Great Cormorant feathers was 127.9°. These data indicate that the physical construction of the feather surface plays an important role for their hydrophobicity, as discussed in the previous article of this issue (Stangier et al. 2023). Therefore, the contact angle of feather fragments was measured before and after degreasing the feathers in pentane. No significant difference (+ 1.8°) in contact angles was found for the Cormorant while a small but significant increase (+ 3.2°) after feather washing was observed for the Duck.

Discussion

The wax ester composition of the uropygial gland of the Great Cormorant comprises > 1000 compounds consisting of a very variable combination of medium- to long-chain acids with up to two methyl groups and slightly longer alcohols with up to three methyl groups, resulting in the enormous number of individual compounds (Tables 1 and S1). This compound diversity in the secretion is indicated by the uncommon appearance of the TIC in the GC/MS analyses (Fig. 2A), visible also in some other birds previously analyzed, e.g., the Sanderling, Calidris alba (Dekker et al. 2000)), or the White-ruffed Manakin, Corapipo altera (Haribal et al. 2009). Unfortunately, in many articles on preen gland compositions, a chromatogram of the unmodified lipids is often missing, leaving it unclear how alcohol and acids are connected in the respective esters, although the individual components were reported (but see e.g., Grieves et al. 2019, Sinninghe Damsté et al. 2000; Dekker et al. 2000; Leclaire et al. 2011, or Rijpstra et al. 2007). Due to the large overlap of compounds, an identification of individual wax esters in the Cormorant was not possible. Nevertheless, derivatizations allowed the identification of wax ester building blocks in both species.

Comparing the identified compounds in Great Cormorant and Muscovy Duck, the variety and the number of different acid and alcohol building blocks as well as the number of wax esters are much higher in the Great Cormorant. The Great Cormorant also has a higher variety of branched compounds and a higher maximum number of methyl branches that can occur in one compound. Dicarboxylic acids and other oxidized fatty acid derivatives were detected as minor constituents besides the major 2,4,6-trimethyloctanoic acid in Muscovy Duck, but are absent in the Great Cormorant. In addition, ethyl 2-phenylpropanoate, not reported before from birds, is a volatile Cormorant gland constituent.

Identification and mass spectrometry

For the analysis of uropygial gland secretions, GC/MS is the method of choice, although it has some limitations in size and structures of the analytes. Although higher order wax types have been reported from some species (Kolattukudy et al. 1987; Sinninghe Damsté et al. 2000; Rijpstra et al. 2007), compounds with a high molecular mass are, if at all, difficult to detect by GC/MS due to their low volatility, leading to decreased peak intensities. Therefore, we looked for a fast preliminary screening method for the molecular gland composition. ASAP-APCI-MS proved to be useful because of its mass range up to 1500 Da, simplicity, relatively inexpensive instrumentation, and the possibility of direct introduction of a secretion sample into the mass spectrometer without any need to solubilize samples. The value of this method can be potentially fully explored when a large number of samples needs to be analyzed, e.g., in comparative analysis of individual’s secretions or species comparisons. The APCI technique allows simple ionization with low fragmentation, fast analysis and a good overview of molecular species within the wax. Similar results were obtained by DI-ESI-MS on a high-end mass spectrometer, although ESI ionization is more selective on structures with polar functional groups and may fail for some potential wax constituents, e.g., hydrocarbons (Gross 2004). The Great Cormorant ASAP-APCI-MS analysis indicated the presence of different compound classes such as esters, fats, and diesters, which were assigned after further analysis. The difference between the two species, a large distribution of wax esters in the Cormorant, and the dominance of two compounds in the Duck is visible. Other compounds such as fats are also detectable. Furthermore, fragmentation of the esters into alkenes and acids visible below m/z 300 can give additional information. A drawback is the occurrence of dimers, both in APCI or ESI mode in the higher mass region. Therefore, further characterization of the secretion is required to verify the structures of the gland constituents.

While saponification gave an overview of the multitude of isomers of the secretion, it does not allow unequivocal structure identification of the wax ester components. The acids and alcohols were identified by combining results from derivatization using methyl, pyridin-3-ylmethyl, and nicotinate esters. Because of the large number of compounds in the samples, a full characterization could not be achieved; some components were only detectable or assignable by one derivatization method. One methyl group is often positioned near the carboxylic or alcohol head group, while additional methyl groups are found toward the tail region. The chain length of the Cormorant’s esters covers a much wider range than that of the Duck that is dominated by hexadecyl and octadecyl 2,4,6-trimethyloctanoate. Additionally, less methyl branches, and no trimethyl-branched building blocks are found in the Cormorant’s secretion. While methyl groups occurring at even-numbered carbons are known to be introduced during fatty acid biosynthesis by methylmalonate building blocks instead of malonate (Buckner and Kolattukudy 1975), it is also noticeable that only the Cormorant produces acids and alcohols with methyl branches located at odd carbon atoms, predominantly C-3. Some of these are paired with a second methyl branch at an odd carbon (Table 1 and Table S1). There are even four alcohols with three methyl branches on odd carbons. This occurrence indicates the presence of an α-oxidation enzymatic machinery that shortens the long carbon chain by one carbon, concomitantly moving the methyl group from even to odd positions. Although such an α-oxidation process has been primarily investigated in terpene-derived phytanic acid degradation (Casteels et al. 2003), this process is likely active also in fatty acid biosynthesis in animals (Goller et al. 2007). The biosynthesis of the alcohols occurs by reduction of the respective fatty-acyl-CoA precursors (Hellenbrand et al. 2011). Interestingly, branching patterns in acids and alcohols match only partly. This suggests a complex fatty acid biosynthesis machinery that partly uncouples pathways to the acid and alcohol parts of the esters. Furthermore, the broad compound distribution implies a rather sloppy biosynthetic machinery, leading to the high diversity in esters by wax synthases (Biester et al. 2012).

Uropygial gland microbiota have been shown to be highly diverse (Whittaker and Theis 2016; Grieves et al. 2021a, b; Talbott et al. 2022), making it not unlikely that bacteria might also be involved in the biosynthesis or modification of uropygial gland secretions. α-Oxidation is also common in fatty acid biosynthesis in bacteria (Bode et al. 2005), which may hint to involvement in the biosynthesis or degradation of the Cormorant wax constituents. Involvement of microorganisms might also be true for the Duck because oxidized compounds such as diesters or oxoesters which might be formed by oxidative degradation of saturated acids or esters by microorganisms.

The two major Duck 2,4,6-trimethyloctanoates have been reported as major constituents of the Muscovy Duck (Priol et al. 1991), as well as the Coscorba Swan, Coscoroba coscoroba, the Common Scoter, Melanitta nigra, and the Tufted Duck, Aythya fuligula, the latter also producing an even greater variety of trimethyl-branched acids. (Zeman 1970; Jacob and Zeman 1972; Andersson and Bertelsen 1975; Jakob and Hoerschelmann 1993; Jacob et al. 1997). All other esters occur in much smaller concentrations, but our detailed analysis also revealed the identities of the minor components in greater variability. The acids with aldehyde and keto groups as well as the dicarboxylic acids were also detected in the Duck secretion. Such compounds have not been reported before to form uropygial gland secretions to the best of our knowledge. They might be formed by oxidative cleavage of unsaturated fatty acids, and may add to a specific odor.

Water repellency

Due to the large differences between the two wax compositions, they represented good candidates to investigate their water repellency. The higher contact angles of the Muscovy Duck wax indicate a higher hydrophobicity. When looking at the contact angles of the feathers, the Muscovy Duck and the Great Cormorant are almost equally effective, with a slightly better performance of the Muscovy Duck. With the wax layer washed away, we expected a decrease in water repellency, but the contact angles did not change significantly in the Great Cormorant, while the Duck showed a small, significant change. This suggests a higher influence of the feather construction (see accompanying article by Stangier et al. 2023) than by the feather wax, although a more in-depth study is required to clarify this point exactly.

The diversity of compounds as well as the lack of dominating compounds in the Great Cormorant secretion may hinder a structured organization within the lipidic phase, e.g., the formation of “compound blocks”, and increases mobility of the fluid by entropic effects. These factors may increase water drift through the lipid layer and may contribute to a reduced formation of micro air bubbles, reducing the buoyancy of the bird in water, as compared to Duck secretion. Nevertheless, the lack of difference in the contact angle measurements between washed and natural feathers seems to contradict influence of the wax layer. The Cormorant feathers were sampled after the birds had been shot in the morning, so we do not know their actual state regarding preening. Furthermore, these effects may be observable only under certain conditions, and the lack of difference cannot be regarded as proof of absence of the influence of the secretion. In addition, the large difference in contact angles between the Duck and the Cormorant secretion underlines the lower water repellency of the Cormorants secretion, likely due to the highly diverse compound mixture. During the diving process, additional factors such as movability of feather structures along each other, similar to greasing movable parts in machines, might also play an important role. Interestingly, washing of the Ducks’ feathers increased the contact angle slightly, indicating also a low influence of the secretion on water repellence. Nevertheless, more contact angle measurements of the feathers under more controlled conditions are needed to clarify the uropygial gland secretion role in water repellency.

The compound diversity found in the Cormorant secretion can also have other or additional reasons than water repellence or greasing. The secretion might also be instrumental for defense against microorganisms or parasites (Møller et al. 2010). A large number of available compounds increases the possibility that some compounds are effective as defence compounds. This strategy is known, e.g., from insects (Hilker and Schulz 1994). Interestingly, the volatile ethyl 2-phenylpropanoate may also support a role in communication of the Cormorants’ uropygial secretion. This compound, not known from birds before, might be used as signal because of the much higher volatility compared to the wax esters. Although we have no indication for its function and only samples from males, this small, unique ester might be used in olfactory communication, as discussed for other species (Grieves et al. 2022). The wing-spreading behavior, by the way, is an ideal posture to release odor due to the large surface area exposed. The wax ester may also play a role in communication, although a specific signal is probably not generated, because of the many compounds. Therefore, one might argue that a non-clear odor avoids easy detection by predators, as discussed for the large diesters found in other birds to avoid nest detection (Salibian and Montalti 2009; Grieves et al. 2022). Nevertheless, the distinctive volatile ester may be specific. The Duck secretion, in contrast, is more specific and may function also as source of oxidized odor components.

General perspective

Our results are mostly in line with previous analyses of feather waxes. With the applied derivatization methods, we were able to characterize the previously not analyzed secretion of the Great Cormorant. Furthermore, we identified additional, previously unknown, wax ester components like certain dimethyl- and trimethyl-branched alcohols and dicarboxylic acids. Monomethyl-branched and few dimethyl-branched alcohols have been reported in other birds such as several pelecaniformes, the Tufted Duck, Common Scoter, the Steamer Duck or the Red Knot (Zeman 1970; Jacob and Zeman 1972; Livezey et al. 1986; Jacob et al. 1997; Dekker et al. 2000). The occurrence of dicarboxylic acids in uropygial gland secretions has not been reported before. These compounds could form diesters with the alcohols reported, with a higher molecular mass, thus avoiding detection by direct GC/MS methods. These diesters may increase water repellency due to their size, and thus hydrophobicity. Diesters have been reported before from birds, but so far only from aliphatic diols or hydroxy acids as key element, being reported in the orders Struthioniformes, Tinamiformes, Apterygiformes, Ciconiiformes, Charadriiformes, and Anatidae (Kolattukudy et al. 1987; Jacob 1992; Reneerkens et al. 2002, 2005; Rijpstra et al. 2007).

Conclusion

We investigated the uropygial gland secretions of the Great Cormorant and of the Muscovy Duck by ASAP-APCI-MS, DI-ESI-MS, GC/MS, and contact angle measurement. The wax composition differed significantly, the Cormorant showing a large variety of carboxylic esters exceeding 1,000 components, while the Duck secretion contained two major esters. APCI-MS showed to be a promising tool to get a fast overview on the composition of an uropygial gland constitution, but needs further MS methods and derivatization for exact structure determination. We expect the Ducks’ secretion to be a stronger water repellent due to their two major compounds and the presence of carboxylic diesters, making ordered structures in the lipidic phase more likely. This interpretation was supported by contact angle measurements that showed a better water repellency of the Duck’s secretion compared to that of the Cormorant, although experiments with washed feathers challenge the importance of the secretion. A deciding factor seems to be the physical structure of the feathers, as discussed in the accompanying article (Stangier et al. 2023).