This is quite similar to a problem that I encounter with just about every article: I get it looking just how I like it, and then the journal says something stupid like "Only documents in our own in house style will be accepted." The problem is that they provide a class file which (they say) must be used, and which (helpfully!) defines a load of commands, which do nothing for the style and only get in the way.
In the past, I'd have to edit each document each time I submitted to a journal (and for some articles, that's a lot, sadly!). Plus there were the various preprint servers that also required different options. Having switched to keeping my articles in a version control system has actually made me less willing to accept this as I want the changes between versions to be substantial rather than just "changed all definitions from \newcommand
to \renewcommand
to override definitions from journal X".
So what I've developed is a meta class file which sets up a load of stuff and then calls the real class file. So at the head of my document I have:
\documentclass[%
a4paper,%
journalX,%
%draft,%
%hopf,%
%journalY,%
%arxiv,%
%article,%
%amsart,%
defaults,%
]%
{myclass}
There, you can see that I submitted it to a journal 'X' and that it was probably previously submitted to journal 'Y'. There are presumably some changes made in draft mode. It happened to be a paper in algebraic topology, so a copy when to the hopf archive, which I've found has some font issues. Plus there are some defaults getting set. The 'a4paper' option (as with any other unknown option) gets passed on to the eventually-selected class file.
Each time I choose a new journal, I have to integrate its class file into this meta class, but that only needs doing once rather than once-per-article. One major saving on this is with organising the title page. All the title information is entered in the same way for all journals, the meta class file sorts it out into the right format for the specific journal (this is actually the major headache when integrating a new journal class file into my system - I could tell you some horror stories!).
The system gets even better when it comes to my lectures. I use beamer now for my lectures and make the slides available to the students. But I don't want them printing out the full set of slides, so I make the handout version available. Plus, to make it easier to maintain a consistent style and cut-and-paste stuff between lectures, I keep all the lectures in one file. When I run LaTeX on it, I want to be able to select: a particular lecture and a particular format (beamer, trans, handout). I do that by putting in to my meta class file a test that looks at the jobname for these details. Then I make suitable symlinks to the main file:
ln -s lectures.tex lecture.handout.2010-08-23.tex
pdflatex lecture.handout.2010-08-23.tex
produces the handout version of the lecture (to be) given on 23rd August 2010.
This solution is, I'm sure, overkill in your situation! However, as I've said, your situation is only the tip of an iceberg.
Update 2012-04-11: Every now and then I get asked about this class file. I've put a page about it on my website which is at: https://github.com/loopspace/myclass. A copy of the file is available there. Please note that: 1) this was written for my use with no expectation that anyone else would ever see it, let alone use it; 2) it was written a long time before this site came into existence - I would probably do it very differently now! (I'd get egreg to write it via a sequence of craftily sculpted questions.)