Is this because you only have the PCR product? Otherwise the easiest would be to repeat the PCR with a proofreading polymerase as they do not add an A at the 3' end.
Apart from this I also guess that the fill up will not work like this.
As far as I know, all DNA polymerases have 5'-3 'dependent tempalte activity. Additionally some have 3'-5 'or 5'-3' exonuclease activity, but none have 3'-5 'polymerase activity. This means that in your case you would have to use activity 3`-5` 'exonuclease to eliminate the outgoing A.
For this you can use the T4 DNA polymerase, as mentioned above. However, due to the thermostability of the Taq polymerase and its terminal transferase activity, it is mandatory to purify your band beforehand, otherwise the 3 'A will be introduced again. This makes the process long and sometimes inefficient, especially if the efficiency of the original PCR is low.
The alternative that has been previously proposed to use a high fidelity polymerase can be more effective, since, on the one hand, these polymerases do not have terminal transferase activity and leave blunt ends in the amplified DNA. On the other hand, most subsequent applications do not require a subsequent purification step of the PCR product. Although do not forget that in the case of using a high fidelity polymerase, it requires a previous step of optimizing the PCR reaction. Everything has its pros and cons ...