Introduction

Freshwater ecosystems are among the most vulnerable to climate change across all latitudes (Boulton 2008; Leuven et al. 2011; Palmer and Räisänen 2002). Over the last three decades, biodiversity declines have been more pronounced in freshwater ecosystems than in marine and terrestrial ecosystems (Collier et al. 2016; Reid et al. 2019). Another notable trend is “climate migration“, indicating a shift in species’ ranges towards the poles (Chen et al. 2011; Poloczanska et al. 2013). The rapid impact of human-mediated climate change on ecosystems aligns with the accelerated expansion of species beyond their native ranges. Human activities, both direct and indirect, seems to have dismantled biogeographic barriers, blurring the regional segregation of Earth’s biota and facilitating invasions (Vitousek et al. 1997). Freshwater environments, among the most commonly inhabited by invasive species, appear particularly susceptible to invasion (Havel et al. 2015; Haubrock et al. 2021; Cuthbert et al. 2021).

The round goby Neogobius melanostomus (Pallas, 1814) is recognized as one of the most detrimental invasive species in Europe (Cerwenka et al. 2023). While shipping serves as an important vector for the expansion of gobiids, socio-economic factors in Europe have also played role in influencing the timing and route of its invasion (Roche et al. 2013). The round goby has become a model species in invasion biology (Cerwenka et al. 2023). However, the factors driving its expansion, current distribution, and potential spread in the future remain unclear. Field observations suggest that the round goby exhibits an affinity for warmer river sections (Jakubčinová et al. 2018), indicating a possible connection between its expansion and elevated temperatures associated with climate change (Harka and Bíró 2007; Kornis et al. 2012; Kornis and Vander Zanden 2010).

The correlation between the temperature effect, time and “speed” of range expansion of the round goby’s distribution is striking. Over the last two decades coinciding with an increase in water temperature, there has been a significant expansion of the round goby (Cerwenka et al. 2023). This expansion is particularly evident in the Danube system, where river regulation for navigational purposes and wide range of other human activities have negatively affected the geo-hydromorphological, hydrological and thermal regime, consequently altering the structure of the ichthyocenoses in the river basin (Harka and Bíró 2007; Bănăduc et al. 2020, 2023) and has led to a northward range shift and an unpredictable westward expansion. With human-mediated global warming are expected to increase air temperatures at a rate of 0.2 °C per decade, climate change predictions suggest a 1.5 °C temperature increase between 2030 and 2052 if the current rate remains continuous (IPCC 2018). However, certain land regions, such as the Arctic, are projected to experience greater warming than the global average (IPCC 2018). Concurrently, the frequency, intensity, and duration of extreme natural events, such as heatwaves, are projected to rise (Meehl and Tebaldi 2004). Under these predicted warming conditions, the round goby may experience benefits, such as faster growth and prolonged reproductive period (Kornis et al. 2012). On the other hand, the occurrence of heatwaves and temperature increases beyond the thermal limits of the species could potentially hinder its further expansion.

The biological response of a species can vary among populations, influenced by their history of temperature exposure (Dittmar et al. 2014) and the latitude at which they are located (Zanette et al. 2006). For aquatic organisms, many of which are ectotherms (Hoffman and Parsons 1997; Woodward 1987), physiological responses play a crucial role as a measure of their reaction to climate change (Mace and Purvis 2008; Berry et al. 2013; Watson et al. 2013), as temperature increase can impact nearly all biological processes of these organisms (Brown et al. 2004; Hammock and Johnson 2014). The round goby has demonstrated high phenotypic plasticity, indicating its tolerance to various environmental factors and adjustability (Kornis et al. 2012; Cerwenka et al. 2023). Under experimental conditions species showed wide thermal tolerance to range from − 1 to + 30 °C (Moskal’kova 1996), with an energetic optimum at 26 °C (Lee and Johnson 2005). Interestingly, the round goby can continue feeding even under 5 °C and is therefore probably capable of invading cold environments as found in the Northern Baltic Sea (Fortes Silva et al. 2019; Adrian-Kalchhauser et al. 2020). However, the limited time exposure and number of affecting factors under experimental study do not necessarily predict species responses under natural conditions, which may be influenced by additional factors. The species exhibits also wide salinity tolerance (Karsiotis et al. 2012).

The observed spatial distribution of the round goby covers a wide temperature range. Considering the current and forecasted temperature increase, the round goby already experienced or will experience temperatures outside its thermal optimum (Kovyrshina and Rudneva 2012). Temperature increase is a well-known pro-oxidant factor, leading to oxidative stress (Lushchak 2011; Birnie-Gauvin et al. 2017). Elevated temperature is documented to promote oxidative stress, characterized by an imbalance in reactive oxygen species (ROS) levels, disrupting cellular metabolism and damaging cellular components (Lushchak 2011; Birnie-Gauvin et al. 2017). The ability to maintain redox balance is crucial for an organism’s resilience to stress (Bagnyukova et al. 2007). This molecular effect can significantly impact the life history of a species, highlighting the importance of developing an effective adaptive response for species survival (Monaghan et al. 2009). In our previous study, we exposed round goby to acute heat shock under laboratory conditions (related to potential heat waves) and evaluated their oxidative stress response. Despite sex difference in reaction, the general pattern observed in gobies indicated that they overcome oxidative stress within 12 h by returning measured parameters to their initial level, suggesting tolerance to thermal stress (Błońska et al. 2021). Given the high plasticity of round goby, we assessed population differences on a wide geographic scale. Our aim was to assess the potential differences in specific physiological parameters related to the oxidative stress response in five round goby populations across a wide latitudinal span. We evaluated level of lipid peroxidation and reduced glutathione as well as activity of glutathione peroxidase and catalase. Considering the variations in thermal conditions, we anticipated that the southernmost population in the native region, already experiencing severe effects of global warming (Kovyrshina and Rudneva 2012), would exhibit elevated levels of the tested parameters. This was expected due to chronic exposure and an increased in steady-state ROS concentration (Lushchak 2011), compared to populations in the invaded range inhabiting milder ecosystems with regard to thermal regime. A similar pattern could be observed in the opposite geographical direction, in the northernmost population, which, on the other hand, experience temperatures much below thermal optimum for a substantial part of the year.

Materials and methods

Study sites and sampling

Round gobies were collected in April–May 2022 from single locations in each of the five countries, covering the species’ native range (Turkey) and non-native ranges (Croatia, Slovakia, Poland, and Finland). All sampling sites were situated along freshwater rivers (Sava, Danube, Radunia, and Bagirganli Rivers; Table 1), except for Finland, where the fish were captured from two sites in the coastal area of Helsinki in the Baltic Sea. However, these sites are characterized by low salinity due to freshwater input and semi-enclosed conditions (measured at 4 PSU at the sites; Table 1). Temperature variations at each sampling point were illustrated using multi-year data. The climate data utilized in Fig. 1 were obtained from WorldClim version 2.1, encompassing climate data for the period 1970–2000, specifically focusing on the maximum temperature of warmest month (Fick and Hijmans 2017). The sampling sites exhibited homogeneity in substrate type, consisting mainly of a mixture of sand and gravel with scattered stones or rip-raps. Only well-established populations (existing for at least 10 years since the initial observation) at each site were sampled (Kornis et al. 2012; Cerwenka et al. 2023). Fish were captured during the reproductive season (from April to mid-May; Table 1). Round gobies were captured using electrofishing (EFGI 650; BSE Specialelektronik Bretschneider, Germany) or fishing rods (Finland only) and euthanized via spinal cord rupture. On-site measurements of total length (TL) to the nearest 1 mm and weight (W) to the nearest 10 mg were taken prior to euthanasia and dissection. Dissection was promptly conducted at the site, with liver and muscle tissue samples removed and then frozen in dry ice. Once delivered to the laboratory, samples were kept in low temperature freezer under −80 °C for further analysis.

Table 1 Water temperature, salinity, and characteristics of round goby Neogobius melanostomus in sampling sites collected in Turkey, Croatia, Slovakia, Poland, and Finland. Median values with inter quartile range (IQR) for total length (TL)
Full size table
Fig. 1
figure 1

Location of the sampling sites where round goby Neogobius melanostomus was collected in April and May 2022

Full size image

All the procedures were performed according to international/national laws, guidelines, and policies. The round goby collection did not necessitate specific approval from an Ethics Committee. Each country followed its national regulations, with additional lease agreements required in Turkey, Croatia, and Poland (references E-67852565–140.03.03–5071302, UP/I-324–01/22–01/5, PGR-W/Z/704/2022, respectively). All procedures were chosen to ensure humane treatment and compliance with ARRIVE guidelines, thus minimizing unnecessary suffering in animals.

Biochemical analysis

Collected samples of liver and tissue were separately homogenised in 100 mmol/L sodium phosphate buffer (pH 7.4, 100 mmol/L KCl, 1 mmol/L Na2-EDTA) with 100 μmol/L PMSF (Phenylmethysulfonyl fluoride) dissolved in ethanol (98%) using a X-120 knife homogenizer (CAT Ingenieurbüro GmbH, Germany), on ice at 3500 rpm for 4 min. All the following results were calculated based on the total protein established using Lowry et al. (1951) method. Each sample was replicated three times.

Oxidative damage was assessed based on the concentration of malondialdehyde (MDA), a byproduct of lipid peroxidation (LPO) (Rice-Evans et al. 1991). Homogenate samples were mixed with a solution containing 20% TCA, 0.6% thiobarbituric acid (TBA) in hydrochloric acid (HCl, 36–38%), and subjected to shaking followed by centrifugation (3000 rpm, 5 min). The resulting supernatant was heated at 100 °C for 15 min, cooled, mixed with n-butanol, and centrifuged again (3000 rpm, 5 min). Spectrophotometric measurements of MDA levels were conducted in the butanol phase at 532 nm and expressed as nmol MDA/mg protein in the homogenate. The molar absorption coefficient (ε) for MDA is 1.56 × 105 M−1 cm−1.

To estimate antioxidant defense mechanisms, we evaluated the concentration of reduced glutathione (GSH), activity of glutathione peroxidase (GPx) and catalase activity (CAT). Ellman’s method (Ellman 1959) with modification was used to determine the level of GSH: 20% trichloroacetic acid (TCA) was added to the homogenates, centrifuged for 10 min at 15 000 rpm under 4 °C. Obtained supernatant with 10 mmol/L 5,5’-dithiobis (2-nitrobenzoic acid) (DTNB) and 500 mmol/L sodium phosphate buffer was incubated for 20 min in darkness. The formation of yellow 5-thio-2-nitrobenzoate ion corresponding to the GSH concentration was measured by its absorbance at 412 nm. The results were calculated and expressed as a μmol GSH/mg protein in homogenate. The molar absorption coefficient (ε) for DTNB is 13.6 × 103 M−1 cm−1.

The activity of glutathione peroxidase (GPx) in homogenates was assessed using tert-butyl peroxide as a substrate according to the method of (Rice-Evans et al. 1991). Glutathione peroxidase was measured in 0.1 mol potassium phosphate buffer (pH 7.00 with 0.1 mmol/L EDTA) with glutathione reductase, 10 mmol/L GSH, 1.5 mmol/L NADPH in 0.1% NaHCO3 and 12 mmol/L BHP. Conversion of NADPH to NADP+ was monitored at 340 nm for 3 min. One unit of GPx activity was defined as the µmol of GSH, which was converted to oxidized glutathione (μmol GSH/min/mg protein), using an extinction coefficient of 6.22 × 103 M−1 cm−1 for NADPH (1 mmol GSH = 0.5 mmol NADPH).

Catalase activity was estimated based on the enzyme ability to decompose hydrogen peroxide (H2O2) (Aebi 1984). Homogenates were combined with a solution containing 54 mmol/L H2O2 in 50 mmol/L potassium phosphate buffer (pH 7.00). The reduction of hydrogen peroxide was monitored over a 1-min period at 240 nm. CAT activity was quantified as the amount of enzyme required to decompose 1 μmol of hydrogen peroxide per minute. CAT activity in the homogenates was expressed as μmol H2O2/min/mg protein.

Data analysis

All the obtained results (GSH, GPx, CAT and LPO) were subjected to Shapiro–Wilk test to determine normality and modified Levene’s test (based on median) to confirm the homogeneity of variance. Due to non-normal distribution of each tested parameter (W = 0.9158 / 0.7999 / 0.8188 / 0.6359 with p < 0.001 for GPx / LPO / GSH / CAT, respectively), median values (inter-quartile range) are provided. The Wilcoxon rank-sum test with continuity correction was used to check the differences between tissues and sexes (within each tissue), while the Kruskal–Wallis rank-sum test to test differences among countries (within each tissue).

Enzyme parameters were analyzed using a Gamma Generalized Linear Mixing Model (GLMM) in the R environment (version 4.3.2; R Core Team 2023) with the glmmTMB package (Brooks et al. 2017). Prior to model fitting, a thorough data exploration was conducted following the guidelines outlined in Ieno and Zuur (2015). This included assessing missing values, identifying outliers in both the response and explanatory variables, checking for homogeneity and zero inflation in the response variable, examining collinearity between explanatory variables, evaluating the balance of categorical variables, and understanding the relationships between the response and explanatory variables. Any missing values were excluded from further analysis. Additionally, sample size data, which exhibited positive skewness and outliers, were removed from the analysis. The response variables CAT, GP, GSH underwent square root transformation to enhance their distribution. The fixed effects included total length, sex, and tissue. To account for the grouping of studies by site, a random intercept ‘country’ was incorporated into the models to introduce a correlation structure, assuming a normal distribution with a variance > 2 year and a mean of zero. The optimal fixed structure of the model was determined through backward selection using AIC (Δ AIC ≤ 2).

We used the cor function of the corplot R-package and Spearman’s rank correlation coefficient to identify correlations among oxidative stress parameters and environmental factors (temperature and salinity).

Results

In total, 98 round goby specimens (10 males and 10 females / site, except Slovakia with M:F = 8:10) were collected from all sites (Table 1). The median total length of collected fish was 91 (80–109) mm, with the largest fish collected in Finland and the smallest in Slovakia. There was no correlation among oxidative stress parameters and environmental factors (rsp ≤ 0.14, p > 0.05), with the exception of the negative significant correlation between GPx and temperature (rsp = -0.16; p = 0.04).

LPO level

The level of oxidative damage in round goby tissues was higher in liver (median 96.62 (70.31–149.97)) than in muscle (6.30 (5.25–8.55)) (W = 7919, p < 0.001; Fig. 2a). Significant differences were observed in liver tissues between sexes (p = 0.01), with higher values in males. Lipid peroxidation also varied among countries (liver: \(\chi\)2 = 21.78, df = 4, p < 0.001). In muscles, there were no significant differences between sexes (p = 0.96) or among countries (\(\chi\)2 = 7.16, df = 4, p = 0.13). The GLMM model for LPO, having the same fixed factors as in GPH, is detailed in Table 2. Notably, muscle tissue and total length exhibited a diminished effect on LPO levels compared to liver. Additionally, the non-native Finland population displayed a positive effect, while the native Turkey population demonstrated a negative effect on LPO levels (Table 2).

Fig. 2
figure 2

Level of (a) lipid peroxidation (LPO) and (b) reduced glutathione (GSH), activity of (c) glutathione peroxidase (GPx) and (d) catalase (CAT) in tissues of round goby Neogobius melanostomus collected in Turkey, Croatia, Slovakia, Poland, and Finland. The bottom and the top of the boxes correspond to the first and third quartiles, thick bands represent the median, and whiskers represent the outlying values (Note: outliers in each parameter were removed)

Full size image
Table 2 Mean estimates for the oxidative stress levels (GPx, GSH, CAT and LPO) of round goby as a function of tissue, TL, sex, salinity and country, modelled using a gamma GLMM
Full size table

GSH level

The level of reduced glutathione differed between tissues (W = 9155, p < 0.001; Fig. 2b), with higher values in the liver (median 0.0294 (0.0226–0.0412)) compared to muscle (0.0031 (0.0020–0.0046)). Females’ liver tissue showed higher level of GSH than males (p = 0.007), however, there was no difference between sexes in muscle (p = 0.232). Differences among countries were pronounced in the liver (\(\chi\)2 = 19.236, df = 4, p < 0.001) only. The best fitted GLMM for GSH, considering tissue type, TL, and sex, revealed that only the tissue type, specifically muscle, had a statistically significant negative influence on GSH levels. However, the effects of TL and sex were not statistically significant (Table 2).

GPx activity

The activity of glutathione peroxidase in the liver tissue of round gobies (median 0.0558 (0.0469–0.0700)) was higher than in muscle tissue (0.0171 (0.0127–0.0238) (W = 9428, p < 0.001; Fig. 2c). The Wilcoxon rank-sum test also revealed a significant difference in GPx activity between males and females in liver tissue (p < 0.001), with higher values in males. However, no significant difference between sexes was found in muscle tissue (p = 0.901). Similar results were obtained in the comparison among countries for liver (\(\chi\)2 = 13.3, df = 4, p = 0.01) and muscle tissues (\(\chi\)2 = 7.0438, df = 4, p = 0.134). The best fitting GLMM, which included tissue type, TL, sex, and salinity with country as a random effect, indicates a statistically significant positive effect of salinity on GPx activity, suggesting that higher salinities are associated with increased GPx activity (Table 2). Additionally, the impact of tissue type and sex on elevated GPx activity is evident in the GLMM results (Table 2).

CAT activity

Catalase activity was different between liver and muscle (W = 9604, p < 0.001; \(\chi\)2 = 19.236, df = 4, p < 0.001; Fig. 2d), with higher values in liver (median 66.48 (53.86–84.86)) compared to muscle (1.54 (0.97–2.59). There was no difference between sexes regarding the enzyme activity (p = 0.907 and p = 0.369 for liver and muscle, respectively). Differences among countries were pronounced in the liver (\(\chi\)2 = 27.534, df = 4, p < 0.001), but not in muscle (\(\chi\)2 = 1.554, df = 4, p = 0.817). The optimal GLMM model for CAT, considering fixed factors such as TL, country, sex, and tissue, is presented in Table 2. Among these factors, only tissue type showed statistical significance, indicating a lower effect of muscle on CAT activity (Table 2).

Discussion

The present study investigated various oxidative stress parameters in round goby across different populations in a wide latitudinal gradient, representing one of the first studies of its kind considering the scale (Birnie-Gauvin et al. 2017). Tissue-specific character of the tested parameters was confirmed, consistent with findings from other studies (Birnie-Gauvin et al. 2017), including those focused on round goby (Błońska et al. 2021, 2023; Kovyrshina and Rudneva 2012). Liver tissue exhibited a more pronounced response compared to muscles, displaying higher values in all tested parameters. The liver is a crucial biosynthetic organ that generates significant amounts of reactive oxygen species (ROS) as byproducts of energy production. It is also the most significant supplier of GSH released into the blood and then supplying other organs. Reduced glutathione (GSH) is not only the first defense line against ROS, but also serves as a cosubstrate in GPx activity and terminating lipid peroxidation (Lushchack 2012). Therefore, this tissue is more responsive compared to muscles and these results are in line with previous observations in round goby (Błońska et al. 2021, 2023).

While differences between males and females were observed, as demonstrated in our previous study (Błońska et al. 2021), no consistent pattern emerged. The samples were collected during the spawning season of the round goby, which may have influenced the obtained results. The round goby is a batch spawner, meaning females can reproduce multiple times during a spawning season (Kornis et al. 2012). Despite substantial investment by females in gonad development, males invest heavily in secondary characteristics such as territoriality, courtship, and brood care. This extended investment over a long spawning period often leads to their death after one reproductive season (Charlebois et al. 1997; Miller 1984). Our previous findings suggested that the increased burden on males during the breeding period may render their tissues less resistant to oxidative stress compared to females (Błońska et al. 2021).

The analysis encompassed samples from five populations from different countries/biogeographical regions, with notable differences observed, particularly in liver tissue. While no uniform trend was evident across the various parameters tested, the majority exhibited lower values in native populations (Turkey), including reduced glutathione level, catalase activity, and oxidative damage (liver tissue). Conversely, the highest values were most often observed in the non-native Croatian population (GSH, CAT, GPx; liver). These findings diverge from our initial predictions considering the consistency and highest values of tested parameters. Depending on the selective pressures under which species evolved, conspecific populations may differ in the way they regulate their redox state (Birnie-Gauvin et al. 2017). Petitjean et al. (2020) studied gudgeon (Gobio gobio) populations from contrasting thermal environments and found differences between them. Individuals from colder environments displayed higher levels of oxidative stress and DNA damage.

Initially, we expected that the native population inhabiting the Black Sea basin and experiencing increased thermic stress, such as heat waves (Kovyrshina and Rudneva 2012), would display enhanced levels of antioxidant defense system and oxidative damage, as temperature is well known factor inducing oxidative stress response (Lushchak 2011; Birnie-Gauvin et al. 2017). However, water temperature did not emerge as a significant factor in any parameter analysed. In mitigating the increase in reactive oxygen species, non-enzymatic GSH and the high molecular mass antioxidant enzyme CAT exhibited the lowest values in the Turkish population as well as oxidative damage, measured via lipid peroxidation. Although samples collected directly from the field are subject to various unquantifiable factors, these observations might suggest a stronger effectiveness of antioxidant defense in the native round goby population, rendering them able to adequately counterbalance oxidative stress.

Hemmer-Brepson et al. (2013) proposed two potential strategies for individuals to cope with increased production of reactive oxygen species, which are not mutually exclusive. Firstly, fish inhabiting warmer environments might enhance their antioxidant activities to mitigate oxidative damage. However, this investing in defenses could potentially reduce reproductive effort due to the energy required. Alternatively, individuals may allocate less energy to antioxidant defenses to prioritize energy for reproduction, potentially resulting in higher oxidative damage and mortality rates. Their study on adult fish (Oryzias latipes) demonstrated the effect of increased temperature on oxidative balance and life history traits. Despite the significant impact on oxidative metabolism, no oxidative damage was observed. Considering the study by Hemmer-Brepson et al. (2013), our research yielded interesting results. The native Turkish population showed no significant effect of temperature on the basal level of evaluated oxidative stress parameters. However, these individuals exhibited the highest relative fecundity (Blońska et al. under publication). Consequently, it appears that neither antioxidant defense nor reproduction is compromised in populations experiencing the highest thermal insult (Kovyrshina and Rudneva 2012).

Non-native populations did not exhibit any latitudinal pattern, however, most of the tested parameters reached highest values in Croatian samples. CAT and GSH showed no difference among non-native populations, while GPx and LPO did. GPx exhibited sensitivity to salinity, with elevated values correlating to higher salinity levels, as observed in round gobies collected in Finland. The species’ northern range in the Baltic Sea features lower temperatures compared to inland European freshwaters and serves as a brackish water habitat with a distinct ionic composition compared to the Caspian and Black Seas (Christensen et al. 2021), which could be reflected in the obtained results. Given that salinity is among one of the most affecting factors in oxidative stress response (Lushchak 2011; Freire et al. 2011; Birnie-Gauvin et al. 2017), it is surprising that only GPx showed significant relationship.

The observed variations in antioxidant parameters among non-native populations suggest that round gobies may exhibit adaptive responses to local environmental conditions. The species exhibit high phenotypic plasticity, especially in the invaded range (Cerwenka et al. 2023), which might very roughly explain observed trends. Christensen et al. (2021) demonstrated that the round goby exhibits undisturbed physiological function and strong resistance to temperature changes, likely aided by significant phenotypic buffering. This adaptive mechanism enables the species to quickly adapt and thrive in new or evolving environments. However, oxidative stress analysis was not included in the present study, and it was suggested as a future research target, which we attempted to cover.

Although we put a lot of effort into standardizing the quality of the collected samples, our study was not devoid of bias. Sampling was conducted during spawning period, and we started from the most southern range moving northward, however, we might have not captured the same moment in each population. Another indication of potential imbalance could be unavoidable handling time as fish collection proceeded to the dissection process. Tissues were collected immediately at the sampling site, however, not all individuals could be processed at the same time. Nevertheless, the same time protocol was implemented at each location by the same person, which should mitigate this effect.

Conclusion

In conclusion, the biochemical analyses conducted in this study shed light on the intricate interplay between round goby physiology and environmental conditions. The observed variations in antioxidant parameters stress the species' adaptability and highlight the importance of considering population-specific responses in ecological assessments. The biochemical analyses results reveal round goby biological-ecological adaptive interdependencies and responses as a synergic cornerstone of factors which at least partly explain the transcontinental spreading success over multiple biogeographic regions of this species including new and evolving ecosystems. While this study provides valuable insights, further research is needed to elucidate the specific environmental factors contributing to the observed variations. Long-term monitoring and experimental studies could enhance our understanding of how round goby respond to environmental stressors and contribute to the ongoing discourse on invasive species ecology.