Dotted lines show the 3 hydrogen bonds that form between the bases cytosine and guanine on adjacent strands of a DNA molecule. Copied from Lehninger, A. Biochemistry, Worth Publishers, 1975; original from Pauling, Linus and Robert B. Corey. 1956. Specific Hydrogen-Bond Formation Between Pyrimidines and Purines in Deoxyribonucleic Acids. Archives of Biochemistry and Biophysics 65:164-181.
A hydrogen atom attached to an oxygen or a nitrogen atom by an ordinary chemical bond (a covalent bond) can also form a second and much weaker bond with an oxygen or a nitrogen atom on a nearby molecule. Such bonds are known as hydrogen bonds. Hydrogen bonds are very common between biologically important molecules. They are the main bonds that hold the 2 complementary strands of a DNA molecule together; the bases adenine and thymine form 2 hydrogen bonds with each other and guanine forms 3 hydrogen bonds with cytosine.
However, if you look at the earliest papers of Francis Crick and James Watson, the duo who discovered the helical structure of DNA, you will notice that their figures show only 2 hydrogen bonds between guanine and cytosine.
This figure from Watson & Crick (1953)1 shows only 2 hydrogen bonds between guanine and cytosine. Missing is the 3rd hydrogen bond between the hydrogen of the -NH2 group of guanine on the left and the oxygen of cytosine on the right. [Red dotted lines mine.]
This omission by Crick and Watson was the subject of an essay in last week's Nature2. The author, Simon Wain-Hobson, points out that it was Linus Pauling who first determined that guanine and cytosine could definitely form 3 hydrogen bonds and that this was published by Pauling and Corey in 1956 (figure and citation above). Watson and Crick had apparently considered but rejected this 3rd hydrogen bond, because they had thought that the hydrogen of the -NH2 group of guanine and the oxygen of cytosine were not aligned properly and that a hydrogen bond between them would be very weak.
Coincidentally, just a couple of weeks ago I was looking thru the collection of Crick's correspondence at the National Library of Medicine when I came upon a letter from Pauling to Crick, dated 18 feb 1963.
Obviously, this was still an issue, at least in Pauling's mind, almost 10 years after Crick and Watson's first papers on the structure of the DNA and 6 years after Pauling and Corey's paper. Pauling was apparently hoping that Crick would include some sort of correction about the number of hydrogen bonds between guanine and cytosine in the published version of his Nobel lecture presented less than 2 months earlier. However, the printed version of Crick's lecture was not directly about the structure of the DNA molecule, but was about the genetic code and he did not mention hydrogen bonds at all.
Curiously, I could not find a response from Crick to Pauling when I searched Crick's correspondence archive between 1963 and 1966.
What makes hydrogen bonds biologically significant is that they are much weaker than ordinary chemical bonds that hold atoms together in molecules. And exactly because of the relative weaknesses of the hydrogen bonds between the base pairs along its 2 strands, a DNA molecule can come apart easily to be copied.
In a sense, then, life on earth exists and evolution has been possible because of those 3 hydrogen bonds between guanine and cytosine and the 2 between adenine and thymine.
Learn more about hydrogen bonds here and here (more advanced).
1. Watson & Crick 1953. Genetical implications of the structure of deoxyrbonucleic acid. Nature 171:964. pdf
2. Simon Wain-Hobson. The third bond. Nature 439:539 (2 February 2006).