Phase Behavior and Hydration of Silk Fibroin Sungkyun Sohn, Helmut H. Strey, ² and Samuel P. Gido* Department of Polymer Science and Engineering, University of Massachusetts, Amherst, Massachusetts 01003 Received September 21, 2003; Revised Manuscript Received March 1, 2004 The osmotic stress method was applied to study the thermodynamics of supramolecular self-assembly phenomena in crystallizable segments of Bombyx mori silkworm silk fibroin. By controlling compositions and phases of silk fibroin solution, the method provided a means for the direct investigation of microscopic and thermodynamic details of these intermolecular interactions in aqueous media. It is apparent that as osmotic pressure increases, silk fibroin molecules are crowded together to form silk I structure and then with further increase in osmotic pressure become an antiparallel -sheet structure, silk II. A partial ternary phase diagram of water-silk fibroin-LiBr was constructed based on the results. The results provide quantitative evidence that the silk I structure must contain water of hydration. The enhanced control over structure and phase behavior using osmotic stress, as embodied in the phase diagram, could potentially be utilized to design a new route for water-based wet spinning of regenerated silk fibroin. Introduction A metastable structure of Bombyx mori silkworm silk fibroin was discovered in the contents of the air-dried silk gland over sixty years ago. This polymorph was named R-silk 1 or silk I 2 and resulted in X-ray patterns indicating a structure considerably different from that of spun silk fiber, which was designated as silk II. It is well-known from the analysis of fiber diffraction patterns that silk II is an antiparallel -sheet structure. 3 However, the metastable nature and the lack of orientation in the silk I samples have complicated the analysis of this structure. Standard methods used to obtain oriented samples for X-ray diffraction studies, such as rolling and drawing, easily transform the silk I into a silk II structure. In addition to mechanical treatments, temperature changes and solvent effects have also been reported to convert metastable silk I into the stable -sheet structure, silk II. 2,4-9 A number of models for the silk I structure have been proposed over the last thirty years based on data from X-ray and electron diffraction, 10 solid-state NMR, 11,12 and molecular simulations. 13,14 Recently, Atkins has proposed a hydrated mesophase model for silk I. 15,16 Our purpose in this contribution is not to try to distinguish among these competing structural models. Instead, we investigate silk phase behavior in aqueous salt solutions, using the osmotic pressure method. This method allows some general conclusions to be drawn concerning the hydration state of the silk I structure. Understanding the nature of the silk I structure is an essential step to understand the natural silk spinning process. In this study, the structural changes of Bombyx mori silkworm silk fibroin were investigated in vitro using osmotic stress. The osmotic stress method provides a means for the direct investigation of the microscopic and thermodynamic details of intermolecular interactions 17,18 of water-soluble biopolymers and has been successfully employed to measure the intermolecular forces in lamellar stacks of lipid bilayers 19 and in hexagonal arrays of semi stiff biopolymers, such as DNA, 20 collagen, 21 and various polysaccharides. 22 The es- sence of this method involves the controlled removal of water molecules from samples in aqueous environments via the application of osmotic pressure from an inert species. The thermodynamic work required to remove the water from the sample (-µ w dN w ) is equivalent to the work (Π dV w ) required to push the molecules, which constitute the structure closer together. Here µ w is the chemical potential of a water molecule, and dN w is the differential change in the number of water molecules involved. The negative sign is added to the work expression for water removal because water molecules are pulled out of the system, and therefore dN w is a negative number. Π is the applied osmotic stress, and dV w is the differential volume change. In practice, the samples are equilibrated, either through a dialysis membrane or across a semipermeable interface, against a large excess of polymer solution whose osmotic pressure is known as a function of concentration. 23 This polymer solution is known as the stressing solution. Charge neutral and highly water-soluble polymers, such as poly(ethylene glycol) (PEG) or dextran, are ‘osmolytes’ well suited for use as osmotic stressing agents in the stressing solution. An osmotic stress experiment for silk fibroin is schematically illustrated in Figure 1. The system is composed of two subphases: stressing solution and silk fibroin solution. As the external osmotic pressure imposed by the stressing solution is applied to the silk subphase, water molecules are exchanged between the two subphases. The interface between the two subphases is permeable to small molecules such as lithium and bromide ions, as well as water molecules, but is not permeable to PEG and silk fibroin due to their high molecular weights. * To whom correspondence should be addressed. ² Present Address: Department of Biomedical Engineering, Stony Brook University, Stony Brook, NY 11794. 751 Biomacromolecules 2004, 5, 751-757 10.1021/bm0343693 CCC: $27.50 © 2004 American Chemical Society Published on Web 04/07/2004