In attempts to produce diffraction-quality crystals of Fabs from two human
IgM cryoglobulins (Pot and Yvo), we observed unexpected interconversions in
crystal morphologies. The Pot Fab crystallized in two forms when polyethyl
ene glycol (PEG) 6,000 was used as the precipitating agent. Broad, relative
ly short rods with hexagonal cross-sections were our crystals of choice. Ho
wever, long, thin rods that were often curvilinear (i.e., "fibril-like" cry
stals) also appeared in profusion. A third form crystallizing in water had
similar but not identical morphology to the fibril-like crystals. In the pr
esence of NaCl, the hexagonal rods were shortened to hexagonal bipyramids a
nd the growth rates of the fibril-like crystals were also greatly reduced.
Under the same conditions, we observed an unusual phenomenon, in which hexa
gonal pyramids grew along the fibril-like crystals, like candy crystals gro
wing on strings. PEG 6,000 was also used to crystallize the Yvo Fab, but th
e early crystals were highly twinned and imperfectly formed. Among the addi
tives tested to improve the quality of Yvo Fab crystals, 1% (v/v) dimethyls
ulfoxide (DMSO) provided the most dramatic effect. Initial clusters of smal
l, irregular crystals were replaced with large crystals suitable for X-ray
analysis over a period of twelve days. (C) 2001 Elsevier Science B.V. All r
ights reserved.