Made an arXiv submission too early; How to minimize the damage?
The answer to your question depends on the exact details of the journal's policy (and it would be good for you to quote them exactly, or give us a link). More precisely, it depends not only on the journal's policy regarding self-archiving, but more importantly on the journal's policy on prior publication.
I know that might sound weird, but these are actually two separate questions. Self-archiving is a question of copyright, a legal issue: it's a question of what rights you retain (and what rights you transfer) according to the agreement you have with the journal's publisher. This copyright agreement is not something that influences the peer review process, it only governs what you agree with the publisher should they actually accept to publish your paper. (The agreement is void if the paper will not be published, obviously.)
Now, the journal's policy on prior publication is not a legal issue, it is a question of scientific (or editorial) policy. It is decided by the journal's editorial board, and should be explicitly spelt out in its guidelines for authors, journal policies or another similarly titled document. Some journals, like those of the American Chemical Society, have a very strict policy on prior publication:
The Journal of the American Chemical Society considers for publication only original work that has not been previously published and is not under consideration for publication elsewhere. When submitting a manuscript, an author should inform the editor of any prior dissemination of the content in print or electronic format. This includes electronic posting of conference presentations, posters, and preprints on institutional repositories and any other Web sites. Any content that has been made publicly available, either in print or electronic format, and that contains a significant amount of new information, if made part of a submitted manuscript, may jeopardize the originality of the submission and may preclude consideration for publication.
Thus, this journal would reject any manuscript that was posted on arXiv (or anywhere else), because it is not considered original material.
Now, what about you? Well, you have to find out what your journal's policy is, and act accordingly. If the journal policy forbids prior publication, then accept that you made a mistake, find a journal that doesn't have such requirements (it depends on fields, but it should not be too difficult), and submit it there.
On the other hand, if the journal policy does not forbid such prior publication, then you only made a smaller mistake: not informing the editor. Thus, if you want to appeal the rejection, you may want to write an apologetic letter to the editor with that information (along with answering the other comments of the reviewers). As you say that this was not the main reason for the rejection, I wouldn't advise appealing though (chances of success are very slim).
While you may not have done anything wrong, violating a journal's preprint publication policy is a big deal. In fact it is such a big deal that if you did something wrong and this was the reason for rejecting the paper, the editor would have told you off in no uncertain terms. More likely the editor got a recommendation of "reject" from one or more reviewers, quickly looked at the reviews and decided that the decision was warranted and forwarded on the decision. The editor's review of the reviews may have been so cursory that he/she didn't even notice the comment about the arXiv preprint.
Update: I clarified the answer, as F'x correctly pointed out that the last part was not related. Sorry, I wrote the answer before having my morning coffee.
Given that the journal allows self-archiving of the preprint on arXiv (as you reported), I don't think this was a reason for rejection. Reviewers may not know all the journal's rules (not professional behavior, but it happens). The editor, on the other hand, is supposed to know the policies. You can still clarify this with that journal's editor.
Now, your "issue" is that you have your article on arXiv and this preprint cannot disappear. There is nothing wrong in having your manuscript there, as posting a preprint on arXiv has the benefit to make your work more visible. You are supposed to gather feedback on your study by posting your preprint there.
Now, there are two cases, according to how your article was published in arXiv.
If you decided to employ arXiv non-exclusive license to distribute, you are in the best position to submit your article to any another journal. If the journal wants "original articles", it does not necessarily mean that it does not welcome a preprint on arXiv. Journals usually make a distinction on this, if any. In any case, you could mention the preprint existence in the cover letter. It is an act of transparency in any case.
If you chose a CC license, be very sure to make a pre-inquiry with the journal editor in chief. As much as I love CC licenses, non Open Access journals may not like that people will be allowed to create derivate works out of your preprint (i.e., much of your article).
If your target journal has a clear policy against arXiv and other non-institutional preprint servers, most likely you have to switch journal for this paper.