Eutrophication, increased temperatures and stratification can lead to massive, filamentous, N₂-fixing cyanobacterial (FNC) blooms in coastal ecosystems with largely unresolved consequences for the mass and energy supply in pelagic and benthic food webs. Mesozooplankton adapt to not top-down controlled FNC blooms by switching diets from phytoplankton to microzooplankton, resulting in a directly quantifiable increase in its trophic position (TP) from 2.0 (herbivore) to as high as 3.0 (carnivore). If this process in mesozooplankton, we call trophic lengthening, was transferred up to higher trophic levels of a food web, a large loss of energy could result in massive declines of fish biomass. We used compound-specific nitrogen stable isotope data of amino acids (CSIA) to estimate and compare the nitrogen (N) sources and TPs of cod and flounder (mesopredators) from areas with influence of FNC blooms (central Baltic Sea) and without it (western Baltic Sea). We tested if FNC-caused trophic lengthening in mesozooplankton is carried over to fish. The TP of cod from the western Baltic, feeding mainly on decapods, was equal to the global mean value (4.1, secondary carnivore). Only cod from the central Baltic, mainly feeding on zooplanktivorous pelagics, had a higher TP (4.8, near-tertiary carnivore), indicating a strong carry-over effect of FNC-caused trophic lengthening from mesozooplankton. In contrast, the TP of molluscivorous flounder (3.2 ± 0.2 in both areas), associated with the benthic food web, was unaffected by trophic lengthening. This suggests that FNC blooms cause a large loss of energy in zooplanktivorous but not in molluscivorous mesopredators. If FNC blooms continue to detour energy at the base of the pelagic food web, the TP of cod will not return to global mean values and the fish stock not recover. Monitoring the TP of key species can identify fundamental changes in ecosystems and provide useful information for resource management.

The fishes were collected during four cruises to the western Baltic Sea (ICES subdivision (SD) 22) and central Baltic Sea (SD 24-25) in January/February 2019 and 2020 and caught by bottom trawl, with a Bacoma cod end (2019) and a TV30 #520 trawl gear (2020) following the standards of the Baltic International Trawl Survey (ICES, 2017). On board, the fish were killed, identified to species level, counted, and measured. The otoliths were extracted and a sample of white muscle tissue was taken from behind the third dorsal fin of 30 cod individuals and from the tail muscle of 21 flounder individuals. All samples were frozen immediately at −20°C for later stable isotope analyses.

The frozen cod und flounder muscle tissue samples were cleaned mechanically and with distilled water to remove surface contaminants, freeze dried (Christ Alpha 1-4), and then ground and homogenized for further processing. For CSIA, ~10 mg of each dried sample was transferred into a heat-resistant borosilicate vial, mixed with 5 ml of 6M HCl solution and 1 ml of internal standard (trans-4 (amino methyl)-cyclohexane carboxylic acid), and hydrolyzed for 24 h at 110°C. The samples were then filtered through cellulose-acetate filters, dried under a nitrogen flow at 50°C, and then derivatized to TFA-isopropyl amino esters (Silfer et al., 1991; Hofmann et al., 2003), which included an additional purification step using a chloroform-phosphate buffer solution (Veuger et al., 2005). The derivatized samples were dissolved in 500 µl of methylene chloride and stored in GC-vials at −20°C until analyzed as described below.

The measurements of the so-called bulk δ¹³C were carried out according to Loick et al. (2007), using Elemental Analyzer Isotope Ratio Mass Spectrometry (EA-IRMS) analyses of the fish muscle tissue samples (Thermo Finnigan Delta Plus + Thermo Flash EA 1112). For this purpose, the dried and powdered fish muscle tissue samples were weighed (~0.5 mg per sample), packed in tin boats and measured by EA-IRMS. Calibration for the total carbon determination was done daily with an acetanilide standard (Merck). All isotope abundances are expressed in δ notation as follows: δ X (‰) = [(R_sample/R_standard) - 1)] * 103, where X is ¹³C, and R is the ¹³C:¹²C ratio. The internal laboratory reference gas for the C-analyses was ultrapure CO₂, which was calibrated against the materials from the International Atomic Energy Agency (IAEA): NBS 22 (mineral oil δ¹³C = −29.74 ‰) and USGS 24 (graphite δ¹³C = −15.99 ‰). In addition, peptones (Merck) were analyzed as in-house standards after every sixth fish sample run. The analytical error for stable isotope ratios indicated by the peptone standards was less than ± 0.2 ‰ for carbon isotopes.

The TFA-isopropyl-derivatized samples were analyzed for their content of δ¹⁵N in 13 AAs; an external standard (16 AA) was also included. The 13 AAs were: alanine (Ala), glycine (Gly), threonine (Thr), serine (Ser), valine (Val), leucine (Leu), isoleucine (Ile), proline (Pro), aspartic acid (Asp), glutamic acid (Glu), phenylalanine (Phe), tyrosine (Tyr), and lysine (Lys). The concentrations of cysteine (Cys), arginine (Arg), and methionine (Met) in the samples were below the qualitative detection limit of the measurement device and could therefore not be determined in the fish muscle tissue samples. During the initial hydrolysis step, glutamine and asparagine were converted into glutamic acid and aspartic acid, respectively, such that they were considered together as Glu + Gln (referred to herein as Glu) and Asp + Asn (Asp) (Brault et al., 2019).

Amino-acid-specific δ¹⁵N values were measured using an isotope ratio mass spectrometer (IRMS, Thermo Finnigan GmbH, MAT 253 MS, Germany) connected via a ConFlo IV interface unit to a gas chromatograph combustion periphery (GC-C, Thermo Scientific Trace 1310 GC, Italy; Thermo Scientific, GC Isolink, Germany). The separation column in the GC consisted of a non-polar column coated with 5% phenyl-polysilphenylenesiloxane (BPX5, 60 m, 0.32 mm inner diameter, film thickness of 1μm; SEG Analytical Science, Ringwood, Victoria, Australia). For each run, 2 µl of sample was injected via a PTV injector in splitless mode. The temperature program was as follows: Initial temperature at 50 °C, heat up at 12 °C / min to 120 °C, hold for 17 min, heat up at 3 °C / min to 180 °C, linger for 10 min, heat up at 5 °C / min to 200 °C, hold for 6 min, heat up to 250 °C at 10 °C / min and hold for 7 min. Each sample was injected and measured at least three times using helium (He) as the carrier gas. The standard deviation from three runs was usually < 1.0 ‰ for all 13 AAs.