In this work, the biological activity of a variety of gold nanoparticles (AuNPs) was investigated. Water soluble AuNPs in a size range from 1–15 nm were synthesized and characterized via UV/Vis spectroscopy, electron microscopy, dynamic light scattering, elemental analysis, infrared (IR) and nuclear magnetic resonance (NMR) spectroscopy. Within a DFG funded cooperation project, the AuNPs were tested for their cytotoxicity against several cell lines including HeLa cells. For phosphine-stabilized AuNPs, stabilized with 3-(diphenylphosphino)-benzenesulfonate sodium salt (TPPMS), a size dependent toxicity with a maximum of toxicity for 1.4 nm sized AuNPs was found. Furthermore, the influence of the ligand binding strength was investigated. It was found that thiol-stabilized AuNPs are considerably less toxic and, in opposite to phosphine-stabilized small AuNPs, do not induce oxidative stress. By electron paramagnetic resonance spectroscopy, it was analyzed with an oxidizable stable radical as substrate if the toxicity of AuNPs correlates with their catalytic activity for oxidation reactions. This was however not the case. With DNA samples from cells that were incubated with AuNPs, the potential of the particles to damage DNA was investigated. For this, the DNA samples were enzymatically digested, and by gas chromatography and subsequent mass spectrometry the concentrations of DNA base oxidation products (lesions) were quantified. Here, no correlation with cytotoxicity was found. For cytotoxic phosphine-stabilized as well as for non-toxic thiol-stabilized AuNPs, no significantly enhanced concentrations of DNA lesions were found, and in contrary, for both cases, the concentrations of two lesions were decreased. A possible correlation with an activation of certain repair enzymes could not yet be proven within this work. It was further investigated whether the cytotoxic gold(I) complex TPPMS-Au(I)-Cl is potentially present as an impurity in AuNP samples and could thus be partially responsible for the cytotoxicity. By 31P-NMR spectroscopy, this question could not be answered unambiguously. A dialysis experiment revealed that the gold complex is not present besides the AuNPs in significant amounts. With 31P-NMR spectroscopy, the equilibrium between associated and dissociated TPPMS ligand molecules of AuNPs in solution could be confirmed. Patch clamp experiments with potassium ion (hERG) channel expressing cells revealed that the small phosphine-stabilized AuNPs block the ion channels irreversibly. Thiol-stabilized AuNPs again showed no effect. In reference experiments in the presence of an excess of TPPMS, the blocking could be inhibited. This result indicates that the blocking species is a gold core which has partially or completely stripped off its TPPMS ligand shell. This hypothesis could be supported by theoretical calculations from one of the cooperation partners. Besides TPPMS, more phosphines were synthesized and used as ligands for AuNPs. A trisulfonate (TPPTS), a carboxylic acid derivative (TPPMC) and a mixed charged ligand with a carboxylic acid and an amine function (TPPMCMA) could successfully be used as AuNP ligands. The resulting species did however not show considerable changes in cytotoxicity. Further, it was attempted to use diphosphine molecules as ligands for 1.4 nm sized AuNPs to investigate the correlation between ligand binding strength and cytotoxicity in more detail. With 3,3‘,3‘‘,3‘‘‘-(ethane-1,2-diylbis-(phosphintriyl))tetrabenzenesulfonate sodium salt (DPPETS), water soluble AuNPs could be synthesized. In a KCN degradation experiment, these showed lower stability than TPPMS-stabilized AuNPs. In differential scanning calorimetry measurements however, a clearly higher thermic stability of the diphosphine-stabilized AuNPs could be verified. A geometric estimation revealed that DPPETS could possibly, due to the chain length of the ethyl group between the two phosphorus atoms, not act as a bidentate ligand. For this reason, more AuNPs with the propyl- and butyl-analogous ligands (DPPPTS and DPPBTS) were synthesized. These showed however an impaired long term stability and were not stable in cell culture medium. By using larger gold colloids (11–13 nm) it could be shown that the synthesized diphosphine ligands are in principle applicable as AuNP ligands. For a potential therapeutic application of toxic AuNPs, 1.4 nm sized AuNPs were stabilized with a mixed ligand shell of TPPMS and TPPMC and further functionalized with (Lys3)-bombesin, an oligo peptide with high affinity towards gastrin-releasing peptide receptors (GRPR) which are expressed by several cancer cells. The functionalization of the AuNPs could be confirmed by IR spectroscopy. Further, cell experiments with GRPR expressing cells revealed that the activity of the bombesin derivative was not inhibited by coupling to the AuNPs.