Here goes my experience from what I've gathered through the years. I'm far from being an expert in the area. I provide support and what I know is from what I hear editors say when providing support or training.
In Brazil researchers are usually keen to reviewing, since it brings status and recognizes their knowledge of a specific subject matter.
It is almost as if it were a regular attribution to review articles, so they don't really complain. This is, from my experience, a common practice worldwide.
No one complains of receiving an article to review. They usually like it, even though they may not say it. It is sometimes considered an honor, as they are providing their field with expert opinions, which can only enhance the quality of the work published.
Providing a clear text, geared to your audience is key to success. The only problem reviewers may have is lack of knowledge or interest in using a software online to do their work (old school) and they are always full of other activities, which may cause delay in responding.
I don't remember the specific e-mail, but you can either edit the locale files, in a text editor that saves as UTF-8, looking for the standard text that is provided in the subject or browse the E-mail section of the Journal Administration for that specific subject. Through OJS's interface you may edit the content of the standard text.
Regarding registration, it will depend on your journal configurations, and how you want to handle user registration. Some journals don't allow reviewer registration, doing this themselves. A small change in the code, that has been ported to OJS 2.2 is the ability to search authors for their reviewing interests, since they are your first list of possible reviewers.
I hope this long text helps....