Characterization of the Polysialylation Status in Ovaries of the Salmonid Fish Coregonus maraena and the Percid Fish Sander lucioperca

In vertebrates, the carbohydrate polymer polysialic acid (polySia) is especially well known for its essential role during neuronal development, regulating the migration and proliferation of neural precursor cells, for instance. Nevertheless, sialic acid polymers seem to be regulatory elements in other physiological systems, such as the reproductive tract. Interestingly, trout fish eggs have polySia, but we know little of its cellular distribution and role during oogenesis. Therefore, we localized α2,8-linked N-acetylneuraminic acid polymers in the ovaries of Coregonus maraena by immunohistochemistry and found that prevalent clusters of oogonia showed polySia signals on their surfaces. Remarkably, the genome of this salmonid fish contains two st8sia2 genes and one st8sia4 gene, that is, three polysialyltransferases. The expression analysis revealed that for st8sia2-r2, 60 times more mRNA was present than st8sia2-r1 and st8sia4. To compare polysialylation status regarding various polySiaT configurations, we performed a comparable analysis in Sander lucioperca. The genome of this perciform fish contains only one st8sia2 and no st8sia4 gene. Here, too, clusters of oogonia showed polysialylated cell surfaces, and we detected high mRNA values for st8sia2. These results suggest that in teleosts, polySia is involved in the cellular processes of oogonia during oogenesis.


Introduction
Sialic acids, a heterogeneous family of acidic monosaccharides, are ubiquitous on the surfaces of eukaryotic cells and are involved in many crucial cellular events [1]. Besides the typical terminal monosialyl residues, sialic acid polymers can also be synthesized in vertebrates. However, in comparison to monosialylation, elongation by further sialic acid residues is mostly restricted to selected cell types and/or distinct points of differentiation [2][3][4]. For instance, proliferating populations of smooth muscle cells are polysialic acid (polySia)-positive during the postnatal development of murine epididymis, whereas at postnatal day 25, no significant polysialylation occurs in these contractile cell clusters [5].

Real-Time Quantitative PCR (RT-qPCR)
The total RNA from C. maraena ovaries was purified with the TriPrep kit (Nucleospin, Macherey-Nagel, Düren, Germany) according to the manufacturer's protocol. S. lucioperca ovaries were homogenized using 2.8 mm ceramic beads (Precellys, VWR/Avantor) at 6000 rpm for 30 s using the Precellys 24 Homogenizer (VWR/Avantor). The whole RNA was extracted using TRIzol Reagent (Life Technologies, Carlsbad, CA, USA) and subsequently purified with the RNeasy Mini Kit (Qiagen, Hilden, Germany). The quantity of RNA was measured with a NanoDrop One c photomometer (Thermo Fisher Scientific, Darmstadt, Germany). In addition, the RNA integrity was verified by electrophoresis on 1%-agarose gels. RNA was reverse transcribed to cDNA using the SensiFAST cDNA Synthesis Kit (Bioline, London, UK) following the manufacturer's protocols.
Real-time quantitative PCR (RT-qPCR) was used to determine the expression of the polySiaTs: the two duplicated genes st8sia2 r-1, st8sia2 r-2, and st8sia4 for C. maraena and st8sia2 for S. lucioperca. For this purpose, the gene-specific primers were designed using software (Pyrosequencing Assay Design software v.1.0.6; Biotage, Uppsala, Sweden) to amplify the fragments between 121 and 226 bp. To identify the st8sia2 sequences in maraena whitefish, the orthologous sequences from rainbow trout Oncorhynchus mykiss (O. mykiss) were blasted against the transcriptome of maraena whitefish [20], deposited at the NCBI Sequence Read Archive (accession code: SRP066290; BioProject PRJNA302355).

Histological Sample Preparation
For histology, tissue samples were cooled on ice and fixed in Bouin (picric acid, formaldehyde, and acetic acid, 15/5/1, v/v/v) for 24 h. In parallel, samples were fixed for 48 h using a Bouin solution with 2% sucrose. Thereafter, the samples were successively dehydrated with a stepwise alcohol gradient and cleared in xylene followed by isopropanol for 27 h in an automatic tissue processor (MT, SLEE, Mainz, Germany). All the samples were embedded after dehydration in a paraffin station (MPS/P2, SLEE, Mainz, Germany). Paraffin sections (5 µm) were cut with a microtome (LEICA RM 2165, Wetzlar, Germany) and transferred to glass slides. After deparaffinization (1 h at 60 • C, rehydrated in different alcohol steps and pressure-cooked in Tris buffer pH 9 for 3 min), the tissues were stained with hematoxylin and eosin (HE). The stained tissue sections were examined with a laser scanning microscope (Carl Zeiss Axio Observer.Z1, Oberkochen, Germany)

Immunohistochemistry Procedure
For immunohistochemistry, the deparaffinized sections were washed with PBS containing 0.2% (w/v) immunoglobulin G (IgG)-free BSA (Carl Roth, Karlsruhe, Germany). After three washes, tissue sections were blocked with PBS containing 2% BSA (w/v) for 1 h at 37 • C. To immunolocalize α2,8-linked polySia, mAb 735 (0.4 µg/mL) was used. To confirm the specific binding of mAb 735, we used two control strategies: (1) omitting the first antibody and (2) degrading the α2,8-linked polySia with endoN overnight at 37 • C (1.34 µg/mL). The tissue sections of both controls were on the same slide as the normally stained sample.
The incubation with the primary antibody was prepared overnight in PBS containing 0.2% (w/v) IgG-free BSA at 4 • C. Subsequently, the tissue sections were washed and incubated for 1 h at room temperature (RT) with a secondary antibody (envision kit+ system-HRP labeled polymer anti-mouse; Dako, Jena, Germany). After washing, the sections were revealed using a peroxidase chromogen for immunohistochemistry (IHC) SIGMAFAST 3,3'-Diaminobenzidine (DAB)-tab (Sigma-Aldrich, St. Louis, MO, USA). The nuclei were stained with hematoxylin for 10 s.

Synteny Analysis
For the synteny (blocks of orthologous genes) and paralogy analyses, the st8sia gene loci and their neighbor genes were localized using NCBI BLAST search, and the detection of paralogous genes was assessed by manual chromosome walking. We chose genes physically close to the gene of interest to identify only the syntenic segment of interest.

Statistical Analysis
The calculated values were analyzed with Graph Pad Prism 8.4.3 software (GraphPad Software, San Diego, CA, USA) using a Mann-Whitney test. The differences were considered statistically significant at p < 0.05.

Polysialylation in Ovary of C. maraena
Kitajima and colleagues found that α2,8-linked polyNeu5Ac was present on salmonid fish eggs [9] and that all polySiaTs were expressed in rainbow trout ovaries [15]. Based on their studies, we investigated the ovaries of C. maraena, which also belongs to the salmonid lineage, for α2,8-linked Neu5Ac polymers with a DP ≥ 8 using mAb 735 [21,26]. To this end, ovarian tissue was lysed, and the protein was enriched. After protein separation by SDS-PAGE and Western blotting against polySia, broad signals between 135 and 180 kDa were visualized, which were absent or reduced after endoN treatment ( Figure 1). EndoN degrades α2,8-linked Neu5Ac polymers [22,27]. The extensive smear is typical for polysialylated proteins and might result from a heterogeneous polySia chain length distribution on polySia carriers [28,29]. In the case of animal 3, the signal was not completely abolished. This could result from an incomplete digest or a background signal. The visualized protein bands, which showed comparable signal intensities in the endoN-treated and untreated aliquots (low kDa area and in animal 3 proteins > 245 kDa) seem to be the result of an unspecific binding of the applied antibodies. In sum, the obtained results strongly suggest that polySia is present in the ovaries of C. maraena. Cells 2020, 9, x FOR PEER REVIEW 5 of 15

Statistical Analysis
The calculated values were analyzed with Graph Pad Prism 8.4.3 software (GraphPad Software, San Diego, CA, USA) using a Mann-Whitney test. The differences were considered statistically significant at p < 0.05.

Polysialylation in Ovary of C. maraena
Kitajima and colleagues found that α2,8-linked polyNeu5Ac was present on salmonid fish eggs [9] and that all polySiaTs were expressed in rainbow trout ovaries [15]. Based on their studies, we investigated the ovaries of C. maraena, which also belongs to the salmonid lineage, for α2,8-linked Neu5Ac polymers with a DP ≥ 8 using mAb 735 [21,26]. To this end, ovarian tissue was lysed, and the protein was enriched. After protein separation by SDS-PAGE and Western blotting against polySia, broad signals between 135 and 180 kDa were visualized, which were absent or reduced after endoN treatment ( Figure 1). EndoN degrades α2,8-linked Neu5Ac polymers [22,27]. The extensive smear is typical for polysialylated proteins and might result from a heterogeneous polySia chain length distribution on polySia carriers [28,29]. In the case of animal 3, the signal was not completely abolished. This could result from an incomplete digest or a background signal. The visualized protein bands, which showed comparable signal intensities in the endoN-treated and untreated aliquots (low kDa area and in animal 3 proteins > 245 kDa) seem to be the result of an unspecific binding of the applied antibodies. In sum, the obtained results strongly suggest that polySia is present in the ovaries of C. maraena.
It should, however, be noted that it cannot be unambiguously denied that the mAb 735 might additionally visualize other polySia species than α2,8-linked Neu5Ac homopolymers. It has been suggested that anti-polySiaNeu5Ac antibodies (e.g., mAb 12E3) may also bind heteropolymers of Neu5Ac and Neu5Gc residues, even if the binding efficiency is lower [30,31]. Furthermore, endoN can also degrade Neu5G-containing polySia in addition to Neu5Ac homopolymers but with lower efficiency [30,31].

Expression Levels of PolySiaT Genes in C. maraena Ovaries
Following the polySia-positive Western blot results, the expression levels of the polySiaT genes were examined using qPCR. In rainbow trout ovaries, three distinct genes are expressed: two st8sia2 variants and one st8sia4 gene [15]. A recent phylogenetic analysis suggests that the two st8sia2 genes It should, however, be noted that it cannot be unambiguously denied that the mAb 735 might additionally visualize other polySia species than α2,8-linked Neu5Ac homopolymers. It has been suggested that anti-polySiaNeu5Ac antibodies (e.g., mAb 12E3) may also bind heteropolymers of Neu5Ac and Neu5Gc residues, even if the binding efficiency is lower [30,31]. Furthermore, endoN can also degrade Neu5G-containing polySia in addition to Neu5Ac homopolymers but with lower efficiency [30,31].

Expression Levels of PolySiaT Genes in C. maraena Ovaries
Following the polySia-positive Western blot results, the expression levels of the polySiaT genes were examined using qPCR. In rainbow trout ovaries, three distinct genes are expressed: two st8sia2 variants and one st8sia4 gene [15]. A recent phylogenetic analysis suggests that the two st8sia2 genes resulted from a fourth duplication of the whole genome in Salmonidae (SGD = R4) after the teleost radiation (TGD = R3) [16] (Figure 2a and Supplemental Figure S1). Thus, the genome of maraena whitefish encodes three polySiaTs, just as in rainbow trout.
resulted from a fourth duplication of the whole genome in Salmonidae (SGD = R4) after the teleost radiation (TGD = R3) [16] (Figure 2a and Supplemental Figure S1). Thus, the genome of maraena whitefish encodes three polySiaTs, just as in rainbow trout.
To determine the relative transcript levels of the polysialyltransferase genes in the ovaries of C. maraena, specific primer pairs for st8sia2-r1, st8sia2-r2, and st8sia4 were applied. The qPCR analysis revealed that the numbers of st8sia2-r1 transcripts (about 46 copies/ng RNA) and st8sia4 transcripts (about 40 copies/ng RNA) were comparably low ( Figure 2b). Remarkably, the number of st8sia2-r2 transcripts was more than 60 times higher (about 3110 copies/ng RNA).
However, it must be mentioned that the ovaries of the nine whitefish were collected from May until December. In most teleost species, reproduction is cyclic, and seasonality attends several physiological alterations [32]. Concerning the polySiaTs, season-dependent expression is possible in rainbow trout ovaries, since in September, the mRNA level of stx-ov (in C. maraena st8sia2-r1) increase and the transcripts of st8sia4 tend to decrease [15]. For this reason, we separately summarized the determined mRNA levels of the C. maraena samples, which were collected during "summer" (from May until August) and "autumn" (from September until December). The mRNA levels for both ST8SiaII enzymes of C. maraena slightly increased, although the changes were not statistically significant ( Figure  2c). In contrast, the st8sia4 transcripts slightly decreased.
In sum, the results demonstrate that ST8SiaII-r2 is the dominantly expressed polysialyltransferase in C. maraena ovaries.  To determine the relative transcript levels of the polysialyltransferase genes in the ovaries of C. maraena, specific primer pairs for st8sia2-r1, st8sia2-r2, and st8sia4 were applied. The qPCR analysis revealed that the numbers of st8sia2-r1 transcripts (about 46 copies/ng RNA) and st8sia4 transcripts (about 40 copies/ng RNA) were comparably low (Figure 2b). Remarkably, the number of st8sia2-r2 transcripts was more than 60 times higher (about 3110 copies/ng RNA).
However, it must be mentioned that the ovaries of the nine whitefish were collected from May until December. In most teleost species, reproduction is cyclic, and seasonality attends several physiological alterations [32]. Concerning the polySiaTs, season-dependent expression is possible in rainbow trout Cells 2020, 9, 2391 7 of 15 ovaries, since in September, the mRNA level of stx-ov (in C. maraena st8sia2-r1) increase and the transcripts of st8sia4 tend to decrease [15]. For this reason, we separately summarized the determined mRNA levels of the C. maraena samples, which were collected during "summer" (from May until August) and "autumn" (from September until December). The mRNA levels for both ST8SiaII enzymes of C. maraena slightly increased, although the changes were not statistically significant (Figure 2c). In contrast, the st8sia4 transcripts slightly decreased.
In sum, the results demonstrate that ST8SiaII-r2 is the dominantly expressed polysialyltransferase in C. maraena ovaries.

Localization of Polysia in Ovaries of Maraena Whitefish
Oogenesis in fish is a dynamic process [33]. Remarkably, oogonia can proliferate throughout the lifetime of fish, and mitosis can take place in mature ovaries, whereas in mammals, its proliferation is restricted to the embryonic and fetal periods [34,35]. Maraena whitefish have asynchronous gonad development, with different cell types being frequently present in close proximity. Figure 3 shows an example of this non-homogeneous ovarial stage organization.

Localization of Polysia in Ovaries of Maraena Whitefish
Oogenesis in fish is a dynamic process [33]. Remarkably, oogonia can proliferate throughout the lifetime of fish, and mitosis can take place in mature ovaries, whereas in mammals, its proliferation is restricted to the embryonic and fetal periods [34,35]. Maraena whitefish have asynchronous gonad development, with different cell types being frequently present in close proximity. Figure 3 shows an example of this non-homogeneous ovarial stage organization. The following cells are displayed: oogonia (Oog); oocytes in primordial growth (Pg) and early and late primary growth (Pg1, Pg2, respectively); cortical alveoli (CA); primary, secondary, and tertiary vitellogenesis (Vtg1, Vtg2, and Vtg3, respectively). This schematic presentation of oogenesis is based on [36,37].
To investigate the localization of polySia in ovaries, tissue sections were treated with mAb 735 followed by differently conjugated secondary antibodies (HRP, Figure 4a; Alexa Fluor 488, Figure 4b).
The obtained results revealed that polySia-positive cell clusters are mainly localized between cell units in previtellogenic stages (Figure 4). The staining did not take place when polySia was degraded with endoN. The following cells are displayed: oogonia (Oog); oocytes in primordial growth (Pg) and early and late primary growth (Pg1, Pg2, respectively); cortical alveoli (CA); primary, secondary, and tertiary vitellogenesis (Vtg1, Vtg2, and Vtg3, respectively). This schematic presentation of oogenesis is based on [36,37]. To investigate the localization of polySia in ovaries, tissue sections were treated with mAb 735 followed by differently conjugated secondary antibodies (HRP, Figure 4a; Alexa Fluor 488, Figure 4b).
The obtained results revealed that polySia-positive cell clusters are mainly localized between cell units in previtellogenic stages (Figure 4). The staining did not take place when polySia was degraded with endoN.  Interestingly, Nakamura et al. found that oogonia in Oryzia latipes ovaries formed clusters of proliferating cells between cells in previtellogenic stages, which can be stained with antibodies against DEAD-box helicase proteins [35]. Based on these studies, we used polyclonal antibodies against DEADbox helicase 4 (DDX4). In parallel, polySia was visualized. The results demonstrate that several DDX4- Interestingly, Nakamura et al. found that oogonia in Oryzia latipes ovaries formed clusters of proliferating cells between cells in previtellogenic stages, which can be stained with antibodies against DEAD-box helicase proteins [35]. Based on these studies, we used polyclonal antibodies against DEAD-box helicase 4 (DDX4). In parallel, polySia was visualized. The results demonstrate that several DDX4-positive cells are polySia positive ( Figure 5) and suggest that oogonia nests might be a source of polySia in the analyzed ovaries of maraena whitefish.

Polysialylation in Ovaries of S. lucioperca
In C. maraena ovaries, with ST8SiaII-r2, only one of the three polySTs (ST8SiaII-r1, ST8SiaII-r2, and ST8SiaIV) was expressed with copy numbers higher than 1000 (Figure 2b). The other two, ST8SiaII-r1 and ST8SiaIV, seemed present at significantly lower levels (less than 100 copy numbers per ng RNA). Thus ST8SiaII-r2 may synthesize the most polySia chains, which are present in clusters of oogonia in C. maraena. However, it is possible that ST8SiaII and IV are cooperatively involved in polysialylation, as hypothesized for rainbow trout ovaries [15]. Interestingly, in other teleost lineages, such as Neoteleostei, ST8SiaIV was lost and only one variant of ST8SiaII is present (Figure 6a and Supplemental Figure S1). Consequently, the availability of various polySiaTs is restricted to ST8siaII.
We conducted synteny and paralogy analyses of the st8sia2-r1 and st8sia2-r2 loci in genomes of Salmonidae. Figure 6c illustrates the presence of a paralogon including these two gene loci and gnrh, rerg, slco3a1a, fam174b, chd2, rgma, lysmd4, and nr2f2a paralogues on chromosomes 1 and 2 in the rainbow trout O. mykiss. This series of genes is well conserved in the pike perch S. lucioperca on scaffold (NW_022173278.1), further suggesting that the C. maraena st8sia2-r1 and st8sia2-r2 are co-orthologues of the S. lucioperca st8sia2 and could have the same function in fish ovaries. The heterogeneous distribution of polySia-positive cell clusters in the tissue sections explains the high level of variance in the polyST expression levels ( Figure 2) and that the intensity of polySia Western blot signals (Figure 1) are not identical in all animals. The tissue parts of ovaries will generally have a different composition of polySia-positive and -negative cells.

Polysialylation in Ovaries of S. lucioperca
In C. maraena ovaries, with ST8SiaII-r2, only one of the three polySTs (ST8SiaII-r1, ST8SiaII-r2, and ST8SiaIV) was expressed with copy numbers higher than 1000 (Figure 2b). The other two, ST8SiaII-r1 and ST8SiaIV, seemed present at significantly lower levels (less than 100 copy numbers per ng RNA). Thus ST8SiaII-r2 may synthesize the most polySia chains, which are present in clusters of oogonia in C. maraena. However, it is possible that ST8SiaII and IV are cooperatively involved in polysialylation, as hypothesized for rainbow trout ovaries [15]. Interestingly, in other teleost lineages, such as Neoteleostei, ST8SiaIV was lost and only one variant of ST8SiaII is present (Figure 6a and Supplemental Figure S1). Consequently, the availability of various polySiaTs is restricted to ST8siaII. more than 10 times higher (about 48,133 copies/ng RNA) than st8sia2-r2, the dominant polysialyltransferase in maraena whitefish ovaries (Figure 6c). This might be a compensatory effect of the loss of ST8SiaIV and the absence of a second ST8SiaII polysialyltransferase. However, it must be mentioned that the environmental/breeding conditions of C. maraena were different compared to S. lucioperca. Thus, a quantitative comparison is hardly possible. Nevertheless, it is obvious that, also in S. lucioperca ovaries, polysialyltransferases are expressed. Subsequently, the polysialylation status of S. lucioperca ovaries was also investigated by immunohistochemistry. Consistent with the results from C. maraena, in S. lucioperca, clustered cell populations were polySia positive (Figure 7). Ovarian tissue sections from these organisms were also stained against polySia and DDX4 in parallel. As illustrated in Figure 8, several oogonia also displayed a co-staining of polySia and DDX4 in S. lucioperca. Thus, despite the absence of ST8SiaIV, polySia is synthesized in these cell populations, likely by the ST8SiaII enzyme.
The heterogeneity of polySia-positive and -negative cells in a cell population is common and depends very often on the state of activation [2][3][4]39]. In the case of neuronal cells, for instance, polySia supports the migration of the cells [40] or influences access to growth factors [41]. Since in the ovaries We conducted synteny and paralogy analyses of the st8sia2-r1 and st8sia2-r2 loci in genomes of Salmonidae. Figure 6c illustrates the presence of a paralogon including these two gene loci and gnrh, rerg, slco3a1a, fam174b, chd2, rgma, lysmd4, and nr2f2a paralogues on chromosomes 1 and 2 in the rainbow trout O. mykiss. This series of genes is well conserved in the pike perch S. lucioperca on scaffold (NW_022173278.1), further suggesting that the C. maraena st8sia2-r1 and st8sia2-r2 are co-orthologues of the S. lucioperca st8sia2 and could have the same function in fish ovaries.
In S. lucioperca ovary samples, the Q-PCR analysis of the st8sia2 expression levels revealed that it is more than 10 times higher (about 48,133 copies/ng RNA) than st8sia2-r2, the dominant polysialyltransferase in maraena whitefish ovaries (Figure 6c). This might be a compensatory effect of the loss of ST8SiaIV and the absence of a second ST8SiaII polysialyltransferase. However, it must be mentioned that the environmental/breeding conditions of C. maraena were different compared to S. lucioperca. Thus, a quantitative comparison is hardly possible. Nevertheless, it is obvious that, also in S. lucioperca ovaries, polysialyltransferases are expressed.
Subsequently, the polysialylation status of S. lucioperca ovaries was also investigated by immunohistochemistry. Consistent with the results from C. maraena, in S. lucioperca, clustered cell populations were polySia positive (Figure 7). Ovarian tissue sections from these organisms were also stained against polySia and DDX4 in parallel. As illustrated in Figure 8, several oogonia also displayed a co-staining of polySia and DDX4 in S. lucioperca. Thus, despite the absence of ST8SiaIV, polySia is synthesized in these cell populations, likely by the ST8SiaII enzyme.
Cells 2020, 9, x FOR PEER REVIEW 11 of 15 numerous remodeling processes take place, comparable polySia-influenced mechanisms are conceivable. However, the obtained results do not enable valid conclusions to be drawn.  The heterogeneity of polySia-positive and -negative cells in a cell population is common and depends very often on the state of activation [2][3][4]39]. In the case of neuronal cells, for instance, polySia supports the migration of the cells [40] or influences access to growth factors [41]. Since in the ovaries numerous remodeling processes take place, comparable polySia-influenced mechanisms are conceivable. However, the obtained results do not enable valid conclusions to be drawn.

Conclusion
A special characteristic of dynamic oogenesis in fish is the possibility of oogonia proliferating throughout the fish lifetime [31]. Intriguingly, our data revealed that in Salmonidae and Percidae, clusters of these cells exhibited polySia on their cell surfaces, although the fish families had different setups of polySiaTs. Whereas the Salmonidae genome contained three genes for polySiaTs, namely st8sia2-r1, st8sia2-r2, and st8sia4, Percidae could only use st8sia2. Interestingly, the st8sia2 gene in Percidae was flanked by genes comparable to st8sia2-r1 and st8sia2-r2 in Salmonidae. In line with this observation, in C. maraena, the mRNA levels of st8sia2-r2 were 60 times higher than the values for st8sia2-r1 and st8sia4. Thus, it seems that polysialylation of oogonia is driven by ST8SiaII variants in both fish families and that polySia is involved in the cellular processes of germ cells during oogenesis in fish. Altogether, these observations further suggest a conserved function for polysialylation found in fish oogonia. Since it is well known that polySia is essentially involved in migration and proliferation processes of neural precursor cells [39], a comparable function in oogonia seems to be likely.

Supplementary Materials:
The following are available online at www.mdpi.com/xxx/s1, Figure S1: Scenario illustrating the evolutionary history of the polysialyltransferase genes in Salmoniformes and in Perciformes;

Conclusions
A special characteristic of dynamic oogenesis in fish is the possibility of oogonia proliferating throughout the fish lifetime [31]. Intriguingly, our data revealed that in Salmonidae and Percidae, clusters of these cells exhibited polySia on their cell surfaces, although the fish families had different setups of polySiaTs. Whereas the Salmonidae genome contained three genes for polySiaTs, namely st8sia2-r1, st8sia2-r2, and st8sia4, Percidae could only use st8sia2. Interestingly, the st8sia2 gene in Percidae was flanked by genes comparable to st8sia2-r1 and st8sia2-r2 in Salmonidae. In line with this observation, in C. maraena, the mRNA levels of st8sia2-r2 were 60 times higher than the values for st8sia2-r1 and st8sia4. Thus, it seems that polysialylation of oogonia is driven by ST8SiaII variants in both fish families and that polySia is involved in the cellular processes of germ cells during oogenesis in fish. Altogether, these observations further suggest a conserved function for polysialylation found in fish oogonia. Since it is well known that polySia is essentially involved in migration and proliferation processes of neural precursor cells [39], a comparable function in oogonia seems to be likely.