Several wild bird and fish species across the Northern Hemisphere have been shown to episodically be thiamine (vitamin B1) deficient, leading to mass-mortality events in offspring. To understand the mechanisms underlying thiamine deficiency, we need a better understanding of the dynamics and somatic allocation of the vitamin. Here, we focus on an ecologically and economically important species, i.e., Atlantic cod (Gadus morhua), which has been suggested to be sensitive to thiamine deficiency. We sampled cod of varying sizes and maturity stages in a system where thiamine deficiency regularly occurs (i.e. Baltic Sea) and compare these with cod from the North Atlantic, where deficiency has not been recorded. Results show that thiamine concentrations were tissue-specific. Concentrations in muscle and liver declined during growth and maturation, whereas concentrations in gonads increased. Of the total thiamine in a female’s body, approximately 70% was allocated to the gonads at the onset of reproduction, suggesting that micronutrients constitute a major investment when spawning. Transketolase activity suggest that livers were saturated with thiamine and there was no evidence of ongoing thiamine deficiency. We show that life cycle and tissue-specific dynamics in thiamine concentrations should be considered when assessing the thiamine status of a species.

Study system & sampling

Cod specimens were sampled in the North Sea, Baltic Proper, and Åland Sea (Fig. 1A). In the North Sea, thiamine deficiency has never been observed, and while cod stocks are declining, cod still show good condition and grow to large sizes (Fig. 1B,C). This is also true for cod in the Åland Sea, whereas most other stocks in the Baltic Sea, including the Baltic Proper, have shown rapid declines, low condition factors, and size truncation. Hence, we used the North Sea and Åland Sea stocks as reference points to compare them to the Baltic Proper stock, which is suffering from the aforementioned symptoms and a suggested thiamine deficiency. Through coalescing the data on thiamine from all three stocks, we seek to better understand the extent of the potential thiamine deficiency as well as the general thiamine dynamics throughout cod development and reproduction.

All sampled cod individuals were caught as part of environmental monitoring programs conducted by the Swedish University of Agricultural Sciences, Department of Aquatic Science (SLU Aqua). A total of 193 specimens (n = 193) were sampled. In the North Sea, 82 individuals were collected through bottom trawling between January and February 2022. Similarly, in the Baltic Proper, 91 individuals were caught between February and March 2022. Detailed information about these expeditions can be found in the cruise reports (Bland and Börjesson 2022, Lövgren and Casini 2022). In the Åland Sea, 20 cod were obtained through cooperation with a local fisherman using stationary deep-water nets that were placed overnight during June 2022 (Heimbrand et al. 2023).

For each cod, the weight, length, and sex (if possible) were recorded. The maturity level was estimated categorically following standardized protocols (Tomkiewicz et al. 2002). For simplicity, we combined categorical maturity stages into three major maturational stages: Preparation (1,2), maturation (3,4), and spawning (5-7). Liver, gonad, and gastrointestinal tract were separated and weighed to the nearest gram. Where possible, approximately 20 g of dorsal muscle, liver, and gonad tissue were vacuum-packed and stored at -80°C for thiamine and transketolase analysis (see below). In cases where the liver or gonad weighed less than 5 g, it was weighed again on land to an accuracy of 0.001 g. Not all individuals caught during the environmental monitoring programs were sampled for this study. We selected individuals from as many size classes as possible to ensure a representative sample of the stocks’ demography. To control for putative variation within stocks and/or environments, we tried to sample specimens from as many sites as possible. Individuals smaller than 14 cm were excluded from the study as their intestinal mass was insufficient for thiamine analysis.

Thiamine analysis

Thiamine was analyzed in muscle, liver, and female gonad tissues according to Brown et al. (1998) with minor modifications. For a detailed description of the process see Todisco et al. (2024). In short, tissues were homogenized and boiled in diluted trichloroacetic acid. The supernatant was washed using a mixture of ethyl acetate and hexane before adding K₃Fe(CN)₆ as a dye. Samples were analysed in a Hitachi Chromaster HPLC system by measuring fluorescence. Three vitamers of thiamine were quantified: free thiamine (TF), thiamine monophosphate (TMP), and thiamine diphosphate (TDP). Concentrations (unit: nmol/g) were normalized for wet and dry weight (see below) and combined to total thiamine (Ttot).

Transketolase

We selected 48 female cod specimens which covered the total size range and liver thiamine concentrations to measure their transketolase activity and latency. To compare the different stocks, we selected 18, 17, and 13 specimens from the North Sea, Baltic Proper, and Åland Sea, respectively. The activity and latency of transketolase was measured in liver tissues using the BCA Protein Assay Kit (ab102536, Abcam; prev. K813-2500/5000, BioVision) and Transketolase Activity Assay Kit (ab273310, Abcam; prev. K2004-100, BioVision). We primarily followed the manufacturer’s protocols, however we extended the protocol of the Transketolase Activity Assay Kit to also measure latency. These modifications were based on published literature measuring transketolase latency (Engelhardt et al. 2020, Jones et al. 2021).

Frozen cod liver tissue (100 – 200 mg) was homogenized in a 2 ml cryovial with 1ml ice-cold Tris buffer and 2 stainless steel balls (Steelball Lysing Matrix) for 1 min (6.5 ms^-^1; FastPrep-24 5G, MP Biomedicals). Samples were then centrifuged at 10,000 x g at 4°C for 15 mins. The supernatant (700 µl) was collected and filtered through a 10 kDA Spin Column. Protein concentrations (µg/ml) were measured according to the BCA Protein Assay Kit. To measure the transketolase activity and latency, lysates were diluted with Tris Buffer to reach protein concentrations ranging between 0.2-0.4 µg/µl. After mixing 30 µl lysate with 270 µl TKT Assay Buffer, 49 µl of the solution were placed in 5 different wells on a 96-microtiter plate (Brand). We added 1 µl of 10 mM TDP (freshly solubilized in TKT Assay Buffer) to two wells to measure the stimulated transketolase activity. To the other three wells 1 µl TKT Assay Buffer was added to measure the basal transketolase activity in two wells and sample background activity in the last well. Reaction Mix (see manufacturer’s protocol, 50 µl) was added to wells prepared for the measurement of transketolase activity. Background Mix (50 µl) was added to sample background controls. Standards, substrate control, and positive control were prepared following the manufacturer’s protocol. We recommend adding a positive control for the latency measurement by including one sample of a specimen known to suffer from thiamine deficiency. This positive control should be prepared and run simultaneously with the actual samples. Using the microplate reader (FLUOstar Omega, BMG Labtech) preheated to 36°C, we measured the fluorescence (excitation: 544 nm; emission: 590 nm) every minute for 60 min. Following the manufacturer’s protocol, we calculated an average basal and stimulated transketolase activity for each specimen. The latency was calculated by dividing the basal activity by stimulated activity, subtracting it from 1 and multiplying it by 100.

Wet and Dry Weights

In cases where enough tissue remained after subsampling for the thiamine analysis, approximately 0.75 g of frozen tissue was weighed to an accuracy of 0.001 g before and after freeze-drying. Using the difference in weight, i.e., the water content, we estimate thiamine concentrations per wet or dry weight.

Statistical analysis

Statistical analyses were performed using R (version 4.4.2). Condition factor, used as a proxy for health, was estimated by extracting the residuals of a fitted model between somatic weight and length. Both variables were log-transformed. Individuals which could not be sexed due to immaturity were excluded from the analyses.

To investigate how stock as well as several phenotypic variables may affect tissue-specific thiamine concentrations (per g wet weight), we constructed an initial linear mixed model (Gaussian distribution) including a three-way interaction between tissue type, maturity stage and sex along with two two-way interactions between tissue type and stock or length (used as a proxy for age) as well as condition factor as a fixed effect and the fish UID as a random effect (App. 1). After evaluating the three-way interaction, separate models were fitted for each tissue type (App. 2). These models included length, stock, condition factor, sex and maturity stage as fixed effects. An interaction effect between sex and maturity stage was tested for and kept in the model if significant. The model investigating thiamine concentrations in gonad tissue lacks sex as a fixed effect as it was only measured in female fish. In all models, thiamine concentrations were log-transformed and length was scaled. To control for non-independence among samples, models were fitted with fish UID as a random effect using the lme4 package. Model fit was evaluated using diagnostic plots of residual distribution with the DHARMa package. P-values were computed based on robust covariance matrix estimation, i.e. Wald tests. To control for potential effects of varying water content of the samples, we rerun all models using thiamine concentrations per dry weight (App. 1 & 2).

To further investigate how the three stocks might be differing in their thiamine status and allocation, we calculated a total thiamine amount for somatic tissues per body weight. This was performed by calculating the total amount of thiamine in the liver tissue (liver weight multiplied by liver thiamine concentration) and other somatic tissues (eviscerated weight multiplied by muscle thiamine concentration used as an estimate) for each specimen. Both were combined and divided by the somatic weight. This provided an estimate of thiamine concentration per gram somatic weight. A model identical to the tissue-specific models (see above) was fitted.

To study thiamine allocation between tissues throughout maturity, we calculated the total amount of thiamine in gonad, liver and muscle (eviscerated weight) by multiplying tissue weight with thiamine concentration for each specimen, i.e. their total thiamine pool. From there, the relative proportion of thiamine allocated to each tissue was estimated. Relative allocation of thiamine was plotted against maturity stages and analyzed visually. Similarly, the relative proportion of the three measured vitamers TF, TMP, and TDP were estimated and plotted against maturity stages.

The effect of length, stock, condition factor and maturity stage on transketolase activity and latency was investigated by fitting models with aforementioned variables as fixed effects.

References

Bland B, Börjesson P. Expeditionsrapport IBTS, januari 2022. Institutionen för akvatiska resurser, Sveriges lantbruksuniversitet (SLU). eISBN: 978-91-576-9972-5; 2022.

Brown SB, Honeyfield DC, Vandenbyllaardt L. Thiamine Analysis in Fish Tissues. American Fisheries Society Symposium. 1998;21:73-81.

Engelhardt J, Frisell O, Gustavsson H, Hansson T, Sjöberg R, Collier TK, et al. Severe thiamine deficiency in eastern Baltic cod (Gadus morhua). PLOS ONE. 2020;15(1):e0227201.

Heimbrand Y, Larsson S, Landfors F, Bergström U. Beståndsstatus för torsk i Ålands hav 2022. Institutionen för akvatiska resurser, Sveriges lantbruksuniversitet (SLU). SLU ID: SLU.Aqua. 2023.5.4-12; 2023.

Jones KS, Parkington DA, Cox LJ, Koulman A. Erythrocyte transketolase activity coefficient (ETKAC) assay protocol for the assessment of thiamine status. Annals of the New York Academy of Sciences. 2021;1498(1):77-84.

Lövgren O, Casini M. Expeditionsrapport BITS, februari/mars 2022. Institutionen för akvatiska resurser, Sveriges lantbruksuniversitet (SLU). eISBN: 978-91-576-9973-2; 2022.

Todisco V, Fridolfsson E, Axén C, Dahlgren E, Ejsmond MJ, Hauber MM, et al. Thiamin dynamics during the adult life cycle of Atlantic salmon (Salmo salar). Journal of Fish Biology. 2024;104(3):807-24.

Tomkiewicz J, Tybjerg L, Holm N, Hansen A, Broberg C, Hansen E. Manual to determine gonadal maturity of Baltic cod. DFU rapport 116-02. Danish Institute for Fisheries Research. 2002.