In the classic paradigm, immunoglobulins represent products of clonal B cell populations, each producing antibodies (Abs) recognizing a single antigen. There is a common belief that IgGs in mammalian biological fluids are monovalent molecules having stable structures and two identical antigen-binding sites. However, human milk IgGs to different antigens undergo extensive half-molecule exchange. In the IgGs pool, only 33 ± 5% and 13 ± 5% of Abs contained light chains exclusively of kappa- or lambda-type, respectively, while 54 ± 10% of the IgGs contained both kappa- and lambda- light chains. All Ab preparations contained different amounts of IgGs of all four subclasses. Interestingly, lambda-IgGs contained an increased amount of IgG2 (87%) and only 3-6% of each of IgG1, IgG3, and IgG4, while kappa-IgGs consisted of comparable (17-32%) amounts of all IgG subtypes. Chimeric kappa-lambda-IgGs consisted of ~74% IgG1, ~16% IgG2, ~5% IgG3 and ~5% IgG4. As the result of the exchange, all IgG fractions eluted from several specific affinity sorbents under the conditions destroying strong immunocomplexes demonstrated high catalytic activities in hydrolysis of ATP, DNA, oligosaccharides, phosphorylation of proteins, lipids, and oligosaccharides. In vitro, an addition of reduced glutathione and milk plasma to two IgG fractions with different affinity for DNA-cellulose led to a transition of 25-60% of Ab of one fraction to the other fraction. Our data are indicative of the possibility of half-molecule exchange between milk IgGs of various subclasses, raised against different antigens (including abzymes), which explains the polyspecificity and cross-reactivity of these IgGs.