Handle the re-org scenario for the proof provider

### Problem description

WIth the `@lodestar/prover` we have the following scenario where we have to update the payloads in case of re-orgs. We need a better way to handle these re-orgs.

https://github.com/ChainSafe/lodestar/blob/08b82af89f83e89ecaa164e8d94e4e296e0ee65e/packages/prover/src/proof_provider/payload_store.ts#L155-L159

### Solution description

One solution could be that `lightclient` exposes a re-org event which tells the base slot where re-org happens and then `prover` should purge and update the payloads from that slot onwards.

### Additional context

This issue was extracted from https://github.com/ChainSafe/lodestar/issues/4706

Start with packages/prover/src/proof_provider/payload_store.ts at lines 155-159, then trace how lightclient and prover currently handle payloads and re-organizations. Compare the proposed re-org event and determine how payloads from the reported base slot onward should be purged and updated; done means re-orged payloads no longer remain stale.

Written by the indexing model from the issue text.