Deoxyhypusine synthase is the key enzyme for synthesizing the hypusine intermediate [Nε-(4-aminobutyl) lysine]. This intermediate plays a crucial role in the hypusination and thereby the activation of eIF5A, which essentially contributes to the survival and proliferation of eukaryotic mammalian cells (Nishimura et al., 2012). To assess the activity of DHPS in mature neural cortical co-cultures at day in vitro 21 (DIV 21), we analyzed eIF5A^Hyp using Western blot and found that the hypusine content in the DHPS-shRNA expressing cultures was reduced by approximately 45% (mean = 56.36; SEM = 13.03), while the levels of eIF5A protein remained constant (Figures 1A–D). This observation may be attributed to the reduced levels of DHPS caused by its knockdown in the neural co-cultures (mean = 52.72; SEM = 6.29). In contrast, the DHPS-overexpressing group showed no effect compared to the control (Figures 1C,D).

Primary DIV21 cortical neurons expressing DHPS knockdown or overexpression constructs were stained for the microtubule-associated protein 2 (MAP2) and used to perform Sholl analysis. Our findings suggest that the inhibition of hypusination resulted in a reduction in the number of dendritic branches both proximally and distally (Figure 1E). Moreover, the total length of dendrites was significantly shortened in DHPS-shRNA expressing neurons compared to the control group (MD = 56.75; SEM = 17.92). In contrast, since DHPS overexpression in neurons does not exhibit significant morphological changes, it might suggest that neurons endogenously possess sufficient amounts of this enzyme.

Modifications in the dendritic tree are not isolated events but have far-reaching impacts on synaptic connectivity, thereby playing a pivotal role in shaping neuronal function and contributing to various cognitive processes (Kaech et al., 2001). Therefore, we next analyzed whether the observed changes in cytoarchitecture were associated with pre- and post-synaptic changes. In DHPS-shRNA expressing neurons, we observed a significant decrease in Homer-1 intensity (mean = 73.94; SEM = 5.8), whereas no detectable alterations were observed in the number of Homer-1 puncta, suggesting a disruption in the composition of the post-synaptic complex (Figure 2A). Similarly, synaptophysin-1-labeled pre-synaptic structures in DHPS-shRNA expressing neurons appeared to be disrupted, as a decrease in fluorescent intensity was also observed here (mean = 70.98; SEM = 11.07; Figure 2A). In contrast, neurons overexpressing DHPS showed changes only in the pre-synaptic compartment, where we observed a decrease in the total number of synaptophysin-1 signals on the one hand (mean = 89.22; SEM = 4.11), but a significant increase of the synaptophysin-1 intensity on the other (mean = 136.20; SEM = 11.01; Figure 2B). In this context, no changes were observed at the post-synapse (Figure 2B). Additionally, we were able to confirm these findings using Western blots and showed that a knockdown of DHPS, combined with a lower eIF5A^Hyp, affects both the post-synaptic and pre-synaptic composition (Homer-1: mean = 62.62; SEM = 12.98 and synaptophysin-1: mean = 73.97; SEM = 9.34), whereas overexpression of the enzyme did not lead to significant changes (Figure 2C). Considering the fact that hypusinated eIF5A plays a crucial role in the translation of polyproline sequences (Shin et al., 2017), we analyzed Shank-2 harboring two proline-rich domains as a respective candidate using Western blot (Figure 2C). The results obtained in this context underline that, on the one hand, downregulation of DHPS results in a significant reduction in the expression of synaptic proteins (mean = 58.36; SEM = 11.81), while, on the other hand, overexpression of DHPS has no consequences for their expression (Figure 2C). In conclusion, our data demonstrate that a deficiency in eIF5A^Hyp caused by downregulated DHPS level negatively impacts synaptic proteostasis. In contrast, overexpression of DHPS neither leads to increased eIF5A^Hyp nor a significant influence on synaptic proteostasis, indicating a saturated capacity of the endogenous DHPS.

The experiments were approved by the local Ethics Commission of the Federal State of Saxony-Anhalt (according to the European Communities Council Directive; 86/609/EEC) and were conducted in accordance with European regulations for ethical care and use of laboratory animals (2010/63/EU). Adult Wistar rats (Rattus norvegicus) were housed in groups of 5–6 animals, under regular 12 h light–dark cycles (lights on 6 a.m.–6 p.m.) with food and water available ad libitum at constant temperature (22 ± 2°C) and air humidity (40–60%). Efforts were made to minimize the number of animals used and to reduce their suffering during experiments.

Primary rat cortical cells were obtained from Wistar rats at embryonic days 18–19 (E18.5). Pregnant rats were anesthetized with isofluran (Baxter GmbH) for 2 min prior to euthanasia. The E18.5 embryos’ brains were dissected and prepared as described by Müller et al. (2015). After mechanical and enzymatical disassociation of the cortices, the cells were seeded in DMEM supplemented with 10% FBS (Gibco), 100 U/mL of penicillin (Gibco), 100 μg/mL of streptomycin (Gibco), and 2 mM L-glutamine (Gibco) onto poly-D-lysine (Gibco) pre-coated well plates (Techno Plastic Product-TPP). After 3 and 24 h, the medium was replaced with Neurobasal™ medium (Gibco) supplemented with 2% B27 (Gibco) and 0.8 mM L-glutamine (Gibco). The cells were fed weekly with Neurobasal™ by 10% of the total volume. The cultures were kept in an incubator constantly at 37°C and 5% CO₂. For Western blot experiments, the cells were seeded in 6-well plates at a density of 300.000 cells/well (TPP). For immunocytochemistry, they were seeded in 12-well plates (TPP) at a density of 40.000 cells/well.

The plasmid pFUGW was a gift from David Baltimore (Addgene plasmid # 14883; http://n2t.net/addgene:14883; RRID:Addgene_14,883); and the plasmids FUGW-H1 and FUGW-H1-Scrambled were gifts from Sally Temple (Addgene plasmid # 25870; http://n2t.net/addgene:25870; RRID:Addgene_25870 and Addgene plasmid # 40625; http://n2t.net/addgene:40625; RRID:Addgene_40,625). For overexpressing DHPS, the entire ORF of the cDNA from rats (GenBank-ID: NM_001004207.1) was cloned into the FUGW vector and expressed under the control of the ubiquitin promoter. For the knockdown of DHPS, sh-oligos were inserted into the FUGW-H1 vector according to the provided cloning protocol, and the shRNA expression cassette was driven by the H1 promoter. The used DHPS-shRNA sequence was: GCTATACGTCCAACCTCAT. For packing the vectors into Lentiviruses, the packing plasmid psPAX2 (a gift from Didier Trono Addgene plasmid # 12260; http://n2t.net/addgene:12260; RRID:Addgene_12,260) and the plasmid for the envelope protein pCMV-VSV-G [a gift from Bob Weinberg (Addgene plasmid # 8454; http://n2t.net/addgene:8454; RRID:Addgene_8,454)] were used. The preparation of lentiviruses was conducted as previously published (Müller et al., 2015). For the knockdown experiments, the FUGW-H1-DHPS-shRNA vector (DHPS-shRNA) was used with the FUGW-H1-Scrambled vector (Scr-shRNA) as a control. For the over-expression experiments, the FUGW vector containing the DHPS ORF (FUGW-DHPS) was used with the empty FUGW vector (FUGW) as a control.

DIV 21 cultures were briefly rinsed with 1X PBS, pH 7.4, containing 1 mM MgCl₂ and 0.1 mM CaCl₂ and fixed with periodate-lysine-paraformaldehyde for 30 min at RT (McLean and Nakane, 1974). Then, the coverslips were washed three times in 1X PBS pH 7.4 for 10 min each, incubated in blocking solution (B-Block) containing 10% (w/v) normal horse serum, 5% (w/v) sucrose, 2% (w/v) bovine serum albumin, and 0.2% (v/v) Triton X-100 in PBS, pH 7.4 for 1 h at RT, and finally mounted in 10% Mowiol (Carl Roth, 0713.1) and stained with primary antibodies diluted in B-Block O.N. at 4°C. The primary antibodies were: rabbit anti-Homer-1 (SySy, 160,003, 1:1000), guinea pig anti-synaptophysin-1 (SySy, 101,004, 1:1000), and mouse anti-MAP2 (Sigma, M4403, 1:1000). On the next day, the samples were washed three times in 1X PBS pH 7.4 for 10 min each and subsequently stained with the indicated Alexa Fluor dye-conjugated secondary antibodies (from Invitrogen and Jackson Immunoresearch Europe) diluted in B-Block for 1 h at RT. Finally, the coverslips were rinsed, mounted with Mowiol (Calbiochem^TM), and imaged using confocal laser scanning microscopy (CLSM; Axio observer Z1 microscope provided with an LSM 710 confocal system; Carl Zeiss group) for synapse evaluation and a fluorescence microscope (Leica DM6 B LED) for Sholl analysis.

Pre- and post-synapses were labeled as described in Chapter 4.4 and evaluated as maximum intensity projection (89.97×89.97 μm–1024 × 1024 pixels) using the ImageJ plugin SynQuant (Wang et al., 2020). Images were taken in Z-stacks with 8-bit and 0.08 μm pixel size using a Plan-Apochromat 63x/numerical aperture 1.4 oil DIC M27 immersion objective, and identical exposure times were used for the quantification of all experimental groups. MAP2 signal was used as a mask to create a surface for quantifying the synapses puncta and their respective mean intensity. For the Sholl analysis, images were captured using objective HC PL Apo 20x/0.80 WD 0.4 mm and measured using the ImageJ plug-in Simple Neurite Tracer (Arshadi et al., 2021).

The cultures were lysed using protein sample buffer (62.5 mM Tris–HCl, pH 6.8, 1% SDS, 10% glycerol, 5% ß-mercaptoethanol, 0.001% bromophenol blue), incubated at 95°C for 5 min, cooled on ice, and stored at –20°C. Protein concentrations were estimated using the amido black assay and adjusted to 1 μg/μl (Schaffner and Weissmann, 1973). For each sample, 10 μg total protein was loaded on 5–20% SDS-polyacrylamide gels and separated at a current of 12 mA/gel, maintained at a constant temperature of 4°C (Laemmli, 1970). Subsequently, proteins were transferred to nitrocellulose (NC) membranes (Protran BA85, 0.22 μm, LI-COR Biosciences) using the Hoefer™ Mighty Small System SE250 (Fisher Scientific) at 200 mA for 90 min at a constant temperature of 4°C. After Ponceau staining, the membranes were blocked for 1 h at RT in a blocking solution (1xTBS, pH 7.4, containing 0.1% Tween 20 and 5% dry milk) and finally incubated O.N. at 4°C with 1xTBS, pH 7.4, containing 0.1% Tween 20 and the respective primary antibody. The next day, the membranes were washed three times with 1xTBS, pH 7.4, containing 0.1% Tween 20. Finally, the membranes were incubated for 1 h at RT in a blocking solution containing peroxidase-conjugated secondary antibodies (Dianova), washed again three times with 1xTBS, pH 7.4, containing 0.1% Tween 20 and finally imaged in a LICOR OdysseyFC (LI-COR) using ECL-solution according to the manufacturer’s protocol (Pierce™ ECL, Thermo Fisher). Primary antibodies used were: mouse/rabbit anti-Actin (CST, 8457S/3700S, 1:2000), mouse anti-DHPS (Santa Cruz, sc-376580, 1:1000), mouse anti-eIF5A (BD, 611977, 1:1000), rabbit anti-Hypusine (Merk EDM Millipore, ABS1064-I-100UL, 1:1000), rabbit anti-Homer-1 (SySy, 160,003, 1:1000), guinea pig anti-synaptophysin-1 (SySy, 101,004, 1:1000), and guinea pig anti-Shank-2 (SySy, 162,204, 1:1000). All images and quantifications were conducted using Image Studio ™ software and Licor Odyssey^FC (LI-COR).

All statistical analyses were performed using GraphPad Prism software (v.8). Repeated measurement two-way ANOVA followed by the Holm–Sídák test was conducted for Sholl analysis in neurons. A two-tailed t-test was applied to quantitative immunoblotting, synaptic puncta, and intensity measurements. The data are shown as mean ± standard error of the mean (SEM). Statistical significance was declared as a p-value of ≤0.05 for all the datasets.