The shaping of the nebula is currently one of the outstanding unsolved problems in planetary nebula (PN) research. Several mechanisms have been proposed, most of which require a binary companion. However, direct evidence for a binary companion is lacking in most PNs. We have addressed this problem by obtaining precise radial velocities of seven bright proto-planetary nebulae (PPNs), objects in transition from the asymptotic giant branch to the PN phases of stellar evolution. These have F-G spectral types and have the advantage over PNs of having more and sharper spectral lines, leading to better precision. Our observations were made in two observing intervals, 1991-1995 and 2007-2010, and we have included in our analysis some additional published and unpublished data. Only one of the PPNs, IRAS 22272+5435, shows a long-term variation that might tentatively be attributed to a binary companion, with P $>$ 22 years, and from this, limiting binary parameters are calculated. Selection effects are also discussed. These results set significant restrictions on the range of possible physical and orbital properties of any binary companions: they have periods greater than 25 years or masses of brown dwarfs or super-Jupiters. While not ruling out the binary hypothesis, it seems fair to say that these results do not support it.