I believe this case study to be unique because of the supposed "breakthrough" advance in HIV research. In this situation, the researchers in question will be scrutinized by the media and public to a greater degree than would occur in a non-health related topic. Certain areas of science, although useful for furthering the advancement of knowledge, do not seem to enter public awareness as deeply as do those affecting medical science or quality-of-life issues. That being said, the inquiry into misconduct should proceed according to the standard policies at the researchers' institution, with the media being notified only if the parties in question are indeed deemed guilty of research fraud.
The first step is obviously for the pharmaceutical company to contact the PI of the lab. This could all be a misunderstanding associated with the lab notebook and protocols used. However, if a formal inquiry is to be made by the company, the immediate supervisor should be notified. In this case, the department chair would be notified, and it would be his or her responsibility to conduct an informal investigation to decide whether a formal hearing should be pursued. Because this is an academic setting, inquiries into misconduct usually have a set format and at this stage, no other organizations need be notified. However, if sufficient need is found for a formal inquiry, then the university and grant source(s) should be informed.
There is an interesting twist in this specific example. HIV is a touchy subject matter, and any research that may lead to a potential therapy tends to receive quite a bit of attention. The research in question penetrated deep into the public consciousness, with major news networks becoming interested in the lab's work. In addition, pharmaceutical companies see a potential income resulting from drug production. If it turns out that the data were indeed falsified, then not only will a retraction in the science journal be necessary, but a public media announcement by the university would also have to be made.
With all the formal inquiries and media attention, someone will have to take responsibility for problems arising from the published work. Who is ultimately responsible for the integrity of the research? Obviously, it should not fall on the two co-authors and collaborators, whose only roles in the project were to provide research materials. It should also not affect the postdoc who edited the paper, for she had no direct input into the actual bench work.
That leaves us with the grad student, mentoring postdoc, and the PI. It seems that the grad student is the initial culprit; he was in a direct position to corrupt the data. His role should be evaluated through the department, and he would most likely be expelled from the program if found guilty. The postdoc--although being in direct contact with the project and grad student--should not assume too much responsibility. It was his role to instruct the grad student in the techniques and theories necessary to complete the project but not to hold his hand through each step.
I believe the ultimate responsibility resides with the PI. There are two main reasons for this. First, she took on the responsibility to train the grad student. Regardless of the size of the laboratory or how busy she is, she accepted this student into the lab and should be responsible for his training--and for the quality of the work that he produces. Last and most importantly, it is the PI's lab. Everything that is produced and written from students and employees in her lab ultimately reflects on her. It is her integrity on the line every time a paper or grant is submitted or a seminar or lecture is given. As senior contributing author and first author, both the grad student and PI are responsible for the research conducted and published and should be held accountable.