The most important difference to non-scientific ones is length. Candidates for positions in business or administrations are generally advised to compose only one or two pages, and to include only personal information, education and job experience plus, perhaps, hobbies insofar as they are not dangerous. In contrast, a professor of the natural sciences in her late forties or her fifties can have a CV of ten pages, and it generally contains much more detailed information.
What follows is a list of sections, with several alternative headings, that one can find in biologists' CVs, obviously informed by colleagues' CVs that I have seen over the years, especially from Germany and the USA. And my own. Some of them can be combined, some of them can be split; there is no definite right or wrong here.
Name, title/position, address, telephone, e-mail
Year, title or diploma obtained, institution, grade
Years, role, institution. So far, so obvious.
For entirely unclear reasons, American colleagues sometimes put stuff here like greek letter memberships (Studentische Verbindungen), having made the shortlist for a position, or even having been offered a position that they then declined, as if any of that were something to be particularly proud of. More fittingly, one could list actual prices and awards, or having been elected to the presidency of a professional organization. Note that American universities have tons of awards for promising students, i.e. small amounts of money for travel or improving their thesis projects, that German universities, for example, simply never thought to provide. Young American postdocs therefore often have ten or so entries in this area while their German counterparts get their first entry when they are 65 and recive a medal or an honorary membership from some organization.
Current and recent research projects/research interests
This should of course be tweaked a bit depending on the position you apply for.
Grants, scholarships, awards
Year, topic, funding body, amount; in this case, "awards" would mean small grants.
Year, topic, conference name.
(Non-invited) papers and posters presented at scientific meetings
Strangely, giving a talk is often called "presenting a paper" even if nothing but the abstract ever gets printed.
List of societies one is a member of, perhaps with information on since when.
Year, university, name of course, perhaps short explanation of scope of course and/or one's role.
Advising students/student supervision
Name of student, university, year of graduation, perhaps thesis topic, perhaps current position of student.
Service on graduate committees
Name of student, university, year of graduation, perhaps thesis topic.
Advising postdoctoral associates
Name, years, perhaps current position of former postdoc.
Time served on editorial boards of scientific journals, time served on panels and councils, time served as president of professional societies, reviewing/refereeing of grant proposals for a funding body, reviewing/refereeing for a scientific journal, conference organization.
Functions in universities or other academic bodies, e.g. as dean, member of review board, etc.
Writing about research for a popular audience, giving public lectures, providing demonstrations etc. for school students or non-professional societies.
Especially early in the career.
Field work experience
Year, country; if relevant, perhaps how many herbarium collections were made.
List of languages, perhaps with level of proficiency.
Can be anything from computer programs one knows to use, lab methods or analysis methods to potentially relevant experiences outside of science, e.g. debating club activities or whatnot.
Can be organized into subsections such as peer-reviewed/journal, book chapters, book reviews, conference proceedings, web publications, manuscripts in review/submitted manuscripts, and even manuscripts in preparation.
Finally, a few general remarks. All the above items that have beginning and end dates are generally presented as tables but without borders. Dates on the left, content on the right, appears to be more popular than the reverse; the same goes for starting the entries with the most recent. More importantly, however, it seems advisable to be consistent about this, as inconsistent presentation of the information will lead to confusion, confusion will lead to anger, and anger may lead to
Scientific CVs are probably not the occasion to get overly edgy. You can certainly go beyond bold font for headings as the only design element, and use blue headings, or white headings in black bars or something like that, but it should still look simple and professional. Obviously unpretentious fonts like Times New Roman, Arial, or Calibri are to be preferred.
For a scientist, it is advisable to keep a mastercopy of the CV as a text document and to continually update it whenever something can be added. Even if you are not actively searching for a job, you may need up to date information on short notice to attach it to a grant application or internal report, and updating your CV after letting it sit for too long can be very tedious. Many colleagues also upload their CV onto their institutional or personal website, so that potential collaboration partners, students or employers can easily find all relevant information, and if it is so publicly available it should not be too outdated. (This complete lack of concern about making all that information public will probably surprise some non-scientists whose employers usually don't even have staff homepages in the first place.)
When you then need to compose a job or grant application, you should go through it and consider what items are most important and what items are superfluous in this particular case. Applying for a position with a strong teaching aspect? Then student supervision and teaching experience get preferential treatment. Applying for a pure research postdoc? Perhaps not even mention the teaching experience, but shuffle your grants, methodological expertise and research interests to the front. That being said, personal information is always expected at the beginning and the list of publications is, despite its importance, is often found at the end of the CV.
When applying for a grant, space is often limited to two or four pages, and the number of publications to be listed is also often restricted to ten or even less. Under these circumstances, it becomes even more crucial to prioritize. In the selection of publications, their relevance to the topic of the grant is of course the most important criterion. If the number of publications is not limited but there is not enough CV space to list them all together with other crucial information, change the heading to "selected publications" (unless the headings are set, of course).
*) Or resume, if you want - I am not too clear on the taxonomy in this case. What I am talking about is what we hand in when we apply for a job in academia.