DoS attacks involve an attack on a third party to prevent someone from using a service, making the third party both an unwilling participant and a victim of the attack. I don't think anyone would consider that free speech, as much as they wouldn't consider it free speech to call in a bomb threat to an auditorium.
However, in this day and age unrelated interests are getting roped into the discussion and expressing their ability to make their own speech. This is better phrased as a Terms of Service attack. If you have agreed to use a service that can revoke your access for reasons you agreed to, then you are in a precarious position and your speech is not easily protected.
Is a boycott "denial of service"? I can't think of how that might be without stretching the terms to meaninglessness.
Is a bomb threat or getting someone fired a form of DoS? I'll agree with you there, and I also consider blacklisting a form of this.
If you ask me whether a company like Visa or First Data has the right to deny access to its services over political speech, I think it's clear that nobody has sued them successfully over this. But I would also argue that the payments market should be better regulated and require service for all legal transactions. At the very least, there should be a public option that is resistant to fraud, one where the payer initiates the transaction with an unshared secret, like with Pix in Brazil. The public option should support the free speech guaranteed by the government.