Removing an insignificant variable from a regression reduces the R Square value excessively.?

I think you're just interpreting the p-value wrongly - 2.74E-27 is extremely small (=0.00000000000000000000000000274), so this variable isn't "insignificant" at all.