Danielli model

Another microscopic technique is to freeze the specimen and then fracture it with a knife. A knife cutting through the frozen specimen splits the membrane down the middle, exposing the inside of the bilayer (fig. 17.13a). If the Davson-Danielli model for membrane structure were correct, the two exposed surfaces would be featureless. However, electron micrographs of metallic casts of such samples reveal surfaces studded with particles of various sizes (fig. 17.13(f)- Additional studies indicate that these particles are proteins that are deeply embedded in the membrane. The particles seen on the inner and outer leaflets of the bilayer usually differ in size and distribution because of an asymmetrical disposition of the proteins across the bilayer. 

During the 1960s, various alternatives to the Davson-Danielli model were proposed. Some investigators abandoned the idea of a phospholipid bilayer and suggested instead that membranes consist of aggregates of lipid-protein complexes. However, in 1972, Jon Singer and Garth Nicol-... 

Since the above facts have been apparent to most biologists, the classical model (Fig. 3) is far from being rejected as a working hypothesis. It should be emphasized that most biologists consider the Danielli model as a basic model in the sense described above, i.e., one embodying only the structural features common to most membranes. Hence a lack of specificity concerning qualitative and quantitative characteristics of components is considered desirable since appropriate qualitative and quantitative specifications of these characteristics should lead to specific membrane models. 

Fig. 3. The Danielli Model. Only half the membrane is represented.

The molecular mechanisms responsible for these properties have not been elucidated (Wenner and Dougherty, 1971) nor has it been established that similar mechanisms are operative in membranes. However, the sheer multiplicity of corresponding properties in modified lipid bilayers and membranes is at least suggestive that the Danielli model, which describes a protein-modified lipid bilayer, could be representative of both structures. 

The Danielli model describes a lipid bilayer sandwiched between two layers of protein. The specific gravity of the hydrocarbon phase of the lipid bilayer (= 0.8) and that of the protein layer (= 1.27) are very different therefore, calculable electron densities along a section through the membrane should show correspondingly large differences. 

The variety of membranes in which the existence of a lamellar system has been demonstrated by X-ray methods suggests tiie general organization predicted by the Danielli model. 

Arguments challenging the validity of the Danielli model appear frequently both in the literature and in discussions at conferences and symposia. Such arguments fall in two main categories those which are derived from observed differences in membranes composition and functions, and those which originate from data seemingly unaccountable by the classical model. 

Since the P/L ratio of many membranes is about 1.22 ( = 55/45), it is clear from Table III that a complement of superstructural protein units, such as described by the Danielli model, would be required to raise the P/L ratio above that obtained from condensed protein monolayer. The percentage of total protein this complement should represent to raise the P/L ratio to 1.22 is shown in brackets in Table III. Note that in the extreme case of the mitochondrial inner membrane (P/L 2), this complement would correspond to 2/3 of the total protein, including ATPase. Corrected for 35% denatured ATPase (Schatz and Saltzgaber, 1969), the proportion of structural protein to lipid is 0.65. This corresponds (Table III) to a condensed pro-... 

Robertson s unit membrane (Robertson, 1960, 1964, 1967) differs from the Danielli model from which it is derived by specifying a narrow range, 70-80 A, for the membrane thickness. Legitimate objections have been raised against this unduly restrictive specification since the width of membranes, as measured by X-ray diffraction, can reach 120 A (Finean et ah, 1966a). 

The position of superstructural units in the Danielli model suggests that their removal could occur without considerable structural imirairment of membranes. However, it also suggests that the function of these units and the specificity of membranes to accept units of a particular type should depend on properties of the primary protein coat to which they are attached. 

The Danielli model does not specify the tertiary structure, size, or shape of the units in the primary coat, nor the nature of bonds holding these units together. Hence allowance is made for variations, some already discovered, in properties of structural protein units, either among different membranes or within the same one (Zahler, 1968). 

Treatment of HeLa cells with trypsin results in the liberation from the surface membrane of carbohydrate and protein. This does not affect the viability of the cells (Weiss, 1958) although as much as 22% reduction of dry mass was observed in experiments where only 20-60% of the total sugars were released (Shen and Ginsburg, 1968). The trypsin-accessible glycoproteins of the HeLa cell membrane do not, therefore, play a major structural role and would be among su-perstructural protein units described by the Danielli model. 

It would appear that the study of micrographs obtained from freeze-etched membranes is fraught with interpretative difficulties and may not provide clear-cut answers except to most general questions, such as whether the Danielli model accounts for the general organization of membranes. The conclusion reached by Staehelin (1968) was that freeze-etched membrane micrographs can best be interpreted on the basis of this classical concept. 

Since it is unlikely that the basic assumption in the above derivation could lead to much overestimation of ratio / it must be concluded that the lipoprotein subunit concept is inherent to the Danielli model. As shown below, this concept, which does not entail any structural change of the classical model, is fundamental in interpreting a large number of data through this model. 

A membrane model should explain convincingly the functions and properties of the structure it represents. In view of the membrane model controversy, it would seem highly appropriate to test the ability of the Danielli model to explain demonstrable dynamic properties of membranes. 

Membranes behave as charged condensers whose capacitance, C, and potential, V, lie around 1 /xF and 50 mV, respectively (Hodgkin and Huxley, 1952). In the Danielli model (Fig. 18a) charges lie in polar regions on both sides of a hydrocarbon-like dielectric layer. 

---