Crosslinking with dimethyl suberimidate reveals a chain of histone octamers in chromatin. The octamer can be isolated free in solution at high ionic strength and pH. The identification of dimers formed by crosslinking reveals two or more contacts of each histone with others within the octamer. The molecular weight (110,000) and pattern of dissociation of the octamer are compatible with the composition (F2A1)2(F3)2(F2A2)g(F2B)2. The four main types of histone associate in pairs in solution: histones F2A1 and F3 constitute a tetramer (1-3) of composition (F2A1)2(F3)2, and histones F2A2 and F2B a mixture of oligomers, possibly dimers or short polymers (1,(4)(5)(6). The finding of a histone tetramer led to a model for chromatin structure (7) (8) containing 0.2 mM NaEDTA, pH 7. F1 was removed by the tRNA exchange method (13) in 40 mM NaCl/3 mM NaEDTA, pH 8.Procedure for Crosslinking. Buffers were 70 mM sodium phosphate for pH 8, ionic strength 0.2; 60 mM NaCl/100 mM sodium borate for pH 9, ionic strength 0.1; and 1.95 M NaCI/100 mM sodium borate for pH 9, ionic strength 2.0. Chromatin in buffer was treated at 23' with Me2Sub, freshly dissolved at 20 mg/ml in buffer, or with dithiobis(succinimidyl propionate), freshly dissolved at 50 mg/ml in dimethylformamide. Solutions were dialyzed against 0.2 mM phenylmethylsulfonyl fluoride/0. 1 mM NaEDTA, pH 7 and freeze-dried for analysis in gels.Sodium Dodecyl Sulfate (NaDodSO4)/Polyacrylamide Gel Electrophoresis. NaDodSO4/5% polyacrylamide tube gels (14), 0.4 cm X 7 cm, were run at 6 mA per gel. NaDodSO4/polyacrylamide slab gels contained 18% acrylamide and were run according to Laemmli (15) with three modifications: the concentration of Tris buffer in the separating gel was increased to 0.75 M; the ratio acrylamide: N,N'-methylene bis-acrylamide was lowered (16) to 30:0.15; and the electrode buffer contained 0.05 M Tris, 0.38 M glycine, and 0.1% NaDodSO4. The slabs were 0.15 cm thick and either 15 cm or 30 cm long. The short gels were run at 30 mA for about 6 hr and the long ones at 4 W or less for about 24 hr, until the bromophenol blue tracking dye had reached the bottom. Tubes and slabs were fixed for at least 1 hr in methanol/acetic acid/water (5:1:5), stained in 0.1% (w/v) Coomassie brilliant blue in the same solvent, and destained by diffusion at 370 in 5% methanol/7.5% acetic acid. Migration was from top to bottom or left to right in all gels and densitometer traces shown.RESULTS AND DISCUSSION Native chromatin Chromatin prepared by a method involving limited nuclease digestion is native, whereas chromatin prepared by conventional methods involving shear is not (12). In the experiments reported here we have either worked directly with the product of nuclease digestion (a mixture of chromatin fragments containing DNA of weight-average size 1600 base pairs, which we refer to as "1600 base-pair chromatin") or with pure chromatin fragments (monomer, dimer, trimer, etc. of the repeating unit) obtained from the mixture by fractionation on a sucrose gradient.Most of ...