Large-scale preparation and in vitro characterization of biologically active human placental (20 and 22K) and pituitary (20K) growth hormones: placental growth hormones have no lactogenic activity in humans.

Expression plasmids containing DNA sequences optimized for expression in Escherichia coli were prepared encoding human pituitary (hGH-N 20K) and placental (hGH-V 20 and 22K) growth hormones. The proteins were expressed in bacteria, refolded and purified to homogeneity by anion-exchange chromatography on Q-Sepharose according to a unique protocol developed for each protein. The yields from 5l of fermentation culture varied between 400 and 700mg of electrophoretically pure, over 95% monomeric protein. Circular dichroism (CD) analysis revealed similarity of the purified hGHs' secondary structure to that of the pituitary hGH-N 22K, except for hGH-V 20K, in which the alpha-helix content was lower. The purified proteins were stable as a 0.1% sterile solution held at pH 10-11 at 4 degrees C for at least one month. All three purified hGH molecules formed a 1:2 complex with hGH receptor extracellular domain (hGHR-ECD), similar to hGH-N 22K. Binding experiments using hGHR-ECD revealed that the differences between the two 22K variants or between the two 20K variants were not significant, except that hGH-V 20K exhibited slightly lower affinity. Somatogenic activity was tested in vitro using FDC-P1 cell lines. Whereas the bioactivity of 22K hGHs and hGH-N 20K in FDC-P1-9D11 cells stably transfected with hGHR was almost equal and two to threefold higher than that of hGH-V 20K, in FDC-P1 3B9 cells stably transfected with rabbit (rb) GHR, the bioactivity of both 20K analogues was significantly (five to ninefold) lower than that of the 22K hormones. The lactogenic activity measured in heterologous assays (Nb2-11C cells and Baf/3 cells stably transfected with the long form of rabbit prolactin receptor) revealed that the activity of hGH-N 20K was close to that of hGH-N 22K in the Baf/3 cells, but 4.5-fold lower in the Nb2 cells. The activity of hGH-V 22K was ninefold less in Nb2 cells and 55-fold less in Baf/3 cells, whereas hGH-V 20K had no lactogenic activity in either bioassay. In contrast, in a homologous lactogenic assay using Baf/3 LP cells stably transfected with hPRLR, the activity of both placental hGHs was nil and the activity of hGH-N 20K was 4.3-fold lower than that of hGH-N 22K. The latter finding raises the question of whether the lack of intrinsic lactogenic activity in the placental hGHs that dominate during pregnancy has any physiological relevance.