Given that 6-bits and 2's complement form representation
therefore we can represent -32 to +31 decimal numbers, therefore beyond this range it cause overflow
Given that subtracting from -15 doesn't cause overflow
let x is your required number,
∴ -15-x can't cause overflow ===> -15-x ≥ -32 and -15-x ≤ 31
1) -15-x ≥ -32
-x ≥ -32+15 ===> -x ≥ -17 ====> x ≤ 17 ---------> For largest take 17
2) -15-x ≤ 31
-x ≤ 31+15 ===> -x ≤ 46 ===> x ≥ -46 ----------> For lowest take -
But Note that -46 is can not represented with 6 bit in 2's complement number...
∴ up to -32 only we can represent ===> x ≥ -32 ----------> For lowest take -32