When restriction enzymes jaggedly cut double stranded DNA it results in so called sticky ends. What is the substance that makes the DNA sticky?
1$\begingroup$ Sticky ends also called cohesive ends, in DNA the ends are sticky means they combine together as glue, because they are complementry. $\endgroup$– user18938Sep 20, 2015 at 16:42
The sticky ends are sticky because they have complementary bases. Typically used restriction enzymes cut the two complementary DNA strands at different spots, generating 'overhang', or sticky ends:
These overhangs allow for perfect base pairing (C with G, A with T), which is the result of hydrogen bonding. Just like water molecules show strong affinity to each other or other molecules with -OH and -H groups (e.g., alcohol), hydrogen bonds between nucleotides glues the ends together.
The principles of base-pairing has been nicely answered by @MadScientist in "Why does A pair with T and G with C?"
What are sticky ends. There is no substance that is attached making the DNA ends "sticky". What has actually happened is an overhang of at least a few nucleotides. Blunt ends are another kind of cut, but have no overhanging residues.
Why are sticky ends sticky? Restriction enzymes usually cut these ends deliberately so that a four nucleotides are overhanging on the 5' end of the double strand. These are complementary to other overhangs and because they are less stable than a bound double strand region, and are able to hydrogen bond easily with complementary bases, they are easier to attach with a ligase.
Below is an image of the base pairs interacting via H bonds.
The "substance" is hydrogen bonds (H-bonds), or rather the potential to form them. Each of the unpaired A/T bases in the sticky ends have the potential to form 2 H-bonds with a complementary T/A, and each of the unpaired G/C bases have the potential to form 3 H-bonds with a complementary C/G.
From the perspective of a biophysicist, H-bonds are often thought of as being the strongest intermolecular interactions. I personally think that it's a little bit lazy and unphysical to see things this way, but the prevailing view in the field is that H-bonds determine specificity in all macromolecular interactions (e.g. the specificity that causes a single DNA strand to bind mostly to complementary DNA strands and poorly to random DNA strands).