I am guessing that adding the simple knot in the line to keep it from slipping through the hole in the PVC spreader
and act an anchor will not be an issue.
I'd probably choose a figure-of-8 knot rather than a simple overhand knot for that purpose to avoid
weakening the wire. Even better would be taking a turn of wire around the pipe so the stress on the
wire is spread out over a larger area, rather than putting all the pull on the knot.
But I'm always tying knots in my antenna wires: usually one at the feedpoint for strain relief (separate
from the electrical connection) and a loop of some sort where the support rope attaches. I use
stranded, insulated wire and treat it as rope in this regard.
Yes, the knots do affect the resonant frequency for a given length of wire - slightly. A loop knot will
often act like a shorted stub of wire, providing some inductance, but not enough to replace the same
amount of wire run straight. The effect depends on exactly where along the wire the knot is placed.
But generally the difference is smaller than the effects of height above ground, wire size, ground
conditions, coupling among the wires in a multi-band antenna, and other things that also affect the
resonant frequency in the real world, and the normal process of tuning a dipole will take them all into
account. (The traditional formulas, including those on the license exam, are just approximations, and
can't be expected to give exact results over the range of dipoles that hams might build.)