Fig. 1. Immunolocalization of ND10-associated proteins after heat shock (HS) of HEp-2 cells. The proteins labeled in each panel are indicated in the upper left and right of the panel in the relevant color. Duration and extent of thermal stress and times of recovery are given at the bottom of each panel. (A) Control cells triple-labeled for Daxx, SC35 and coilin. (B) Control cells triple-labeled for Daxx heat-shock factor (HSF1) and nucleoli/nuclear envelope (NO/NE, fibrillarin and lamin B by a mixture of two human autoantibodies). (C) Control cells labeled for PML and Daxx. Both proteins colocalize to ND10 but some staining is found throughout the nucleus, excluding the nucleolus. (D) Control cells labeled for Sp100 and SUMO-1. Most Sp100-positive sites are also positive for SUMO-1. (E) Same as A but after 12 minutes of heat shock. Daxx is dispersed but not SC35 or coilin. (F) Same as B but after 12 minutes of heat shock. Daxx is dispersed but not the nucleolus and nuclear envelope. There is a slight repositioning of HSF1. (G) Same as C but exposed to 42°C for 12 minutes. More PML- and Sp100-containing sites are detected. Not all the new small sites contain both proteins. (H) Same as D but exposed to 42°C for 12 minutes. SUMO-1 is absent from all PML aggregates. (I) Same as A but exposed to 42°C for 1 hour. Little or no Daxx is seen in the PML aggregates but SC35 and coilin are maintained in their normal distribution. (J) Same as B but exposed to 42°C for 30 hours. Daxx is totally dispersed and fibrillarin has left the nucleolus. HSF1 has aggregated in specific domains. (K) Same as C but exposed to 42°C for 1 hour. Daxx is absent from PML aggregates. (L) Same as D but exposed to 42°C for 1 hour. Neither SUMO-1 nor Sp100 is present in concentrated aggregates, although Sp100 is still present in a `sandy' distribution in some cells (top). (M) Cells exposed to 42°C for 1 hour were allowed to recover for 2 hours at 37°C. SUMO-1 starts to reappear in some PML aggregates. (N) Different cells within the same experiment as in M but showing the variation between no recovery of Daxx at ND10 and localized return of Daxx to a subset of ND10, demonstrating the high diversity of the recovery process. (O) Apparent total segregation of Daxx and PML to a few sites after 5 hours of recovery from HS.