Following a recent low-temperature crystal structure analysis of the native xylanase from Penicillium simplicissimum [Schmidt et al. (1998) Protein Sci. 7, 2081-2088], where an array of glycerol molecules, diffused into the crystal during soaking in a cryoprotectant, was observed within the active-site cleft, we utilized monomeric xylose as well as a variety of linear (Xn, n = 2 to 5) and branched xylooligomers at high concentrations (typically 20% w/v) as cryoprotectant for low-temperature crystallographic experiments. Binding of the glycosidic moiety (or its hydrolysis products) to the enzyme's active-site cleft was observed after as little as 30 s soaking of a native enzyme crystal. The use of a substrate or substrate analogue as cryoprotectant therefore suggests itself as a simple and widely applicable alternative to the use of crystallographic flow-cells for substrate-saturation experiments. Short-chain xylooligomers, i.e., xylobiose (X2) and xylotriose (X3), were found to bind to the active-site cleft with its reducing end hydrogen-bonded to the catalytic acid-base catalyst Glu132. Xylotetraose (X4) and -pentaose (X5) had apparently been cleaved during the soaking time into a xylotriose plus a monomeric (X4) or dimeric (X5) sugar. While the trimeric hydrolysis product was always found to bind in the same way as xylotriose, the monomer or dimer yielded only weak and diffuse electron density within the xylan-binding cleft, at the opposite side of the active center. This suggests that the two catalytic residues divide the binding cleft into a "substrate recognition area" (from the active site toward the nonreducing end of a bound xylan chain), with strong and specific xylan binding and a "product release area" with considerably weaker and less specific binding. The size of the substrate recognition area (3-4 subsites for sugar rings) explains enzyme kinetic data, according to which short oligomers (X2 and X3) bind to the enzyme without being hydrolyzed.