Extremophilic organisms represent unexplored resource of biotechnologically exploitable enzymes that can work in harsh conditions, e.g., low or high temperatures, presence of organic solvents and inorganic salts [1]. Novel haloalkane dehalogenase DmxA originates from Marinobacter sp. ELB17 isolated form Antarctic lake [2]. Despite its psychrophilic origin, the enzyme exhibits unusually high melting temperature (Tm = 65.9 ± 0.1 °C) among so far characterized wild-type haloalkane dehalogenases. DmxA also possesses broad substrate specificity and high enantioselectivity towards β-substituted bromoalkanes (E > 100) and brominated esters (E > 200). Gel permeation chromatography, native gel electrophoresis and X-ray crystallography revealed that the enzyme exists in the solution in monomer-dimer equilibrium due to cysteine bridge formation between individual subunits of DmxA dimer. Since cysteine bridges are often responsible for protein stability, we examined its role in DmxA stability by testing enzyme activity and stability under reducing conditions and also by site directed mutagenesis. Although we successfully disrupted the bridge, thermal stability and catalytic properties of the enzyme remained unaffected. DmxA also differs from other family members by one of two halide-stabilizing residues (Gln instead of Asn). According to the orientation of Gln in the enzyme structure, it was found that the residue can contribute to enzyme stability by extra hydrogen interactions with surrounding amino acids. In order to investigate the impact of unusual halide stabilizing residue on DmxA stability, unique Gln residue was replaced by commonly present Asn. While introduced mutation significantly changed substrate specificity and catalytic activity, it had a negligible effect on DmxA stability. We further examined accessibility of the access tunnels and hydrophobic pocket of the enzyme. Molecular dynamics followed by CAVER [3] analysis revealed that DmxA has very narrow tunnels (with average bottleneck radius 1.0 to 1.2 Å) that are mostly closed during the simulations. In order to investigate the hypothesis that narrow tunnels are responsible for high thermostability of the enzyme, two bulky residues located at the tunnel mouth were replaced by smaller residues. The resulting mutant exhibited 9 °C lower thermostability than the wild-type enzyme. Simultaneously, introduced mutations in the tunnel affect enzyme activity, substrate specificity and enantioselectivity. These results suggest that narrow tunnels in DmxA contribute to its paradoxical stability.