In an experiment, an activator cannot bind its enhancer because the enhancer motif is mutated. When the activator's activation domain is artificially tethered near the promoter, transcription partly returns. Which model best explains the rescue?