DNA polymerases have a very tight active site, with very little space for errors to be incorporated. Due to Watson-Crick canonical base pairing theory stating a purine must pair with a pyrimidine to maintain order in the ordered structure of DNA molecules, distortions would occur in the sugar-phosphate backbone if two purines or two pyrimidines (eg- A-G, T-C) bound together. This would stall the polymerase, preventing further replication. It is possible for polymerases to remove the erroneous nucleotides, using the exo domain. Hoogsteen base pairing is also possible, however this will still be between canonical nucleotides (A-T, G-C).
The steric gate is formed by the finger domain. For full catalytic activity, the finger and thumb domain must close fully around the DNA. If this does not happen, due to misaligned DNA or another issue, the polymerase will not translocate quickly and will detach from the DNA. This also allows the polymerase to distinguish between dNTPs and rNTPs, ensuring the correct monomers are incorporated – especially thymine and uracil, as these nucleotides are specific to DNA and RNA (respectively). This structure is found in most DNA polymerases.
A phenylalanine residue is found in the steric gate, allowing selectivity for specific sugars (in the case of DNA polymerase, deoxyribose). This phe residue ensures that only the correct nucleotides are incorporated into the DNA molecule.