Date Published: January 26, 2017
Publisher: Public Library of Science
Author(s): Justin Sigley, John Jarzen, Karin Scarpinato, Martin Guthold, Tracey Pu, Daniel Nelli, Josiah Low, Keith Bonin, Thomas Abraham.
The interior of cells is a highly complex medium, containing numerous organelles, a matrix of different fibers and a viscous, aqueous fluid of proteins and small molecules. The interior of cells is also a highly dynamic medium, in which many components move, either by active transport or passive diffusion. The mobility and localization of proteins inside cells can provide important insights into protein function and also general cellular properties, such as viscosity. Neoplastic transformation affects numerous cellular properties, and our goal was to investigate the diffusional and binding behavior of the important mismatch repair (MMR) protein MSH2 in live human cells at various stages of neoplastic transformation. Toward this end, noncancerous, immortal, tumorigenic, and metastatic mammary epithelial cells were transfected with EGFP and EGFP-tagged MSH2. MSH2 forms two MMR proteins (MutSα and MutSβ) and we assume MSH2 is in the complex MutSα, though our results are similar in either case. Unlike the MutS complexes that bind to nuclear DNA, EGFP diffuses freely. EGFP and MutSα-EGFP diffusion coefficients were determined in the cytoplasm and nucleus of each cell type using fluorescence recovery after photobleaching. Diffusion coefficients were 14–24 μm2/s for EGFP and 3–7 μm2/s for MutSα-EGFP. EGFP diffusion increased in going from noncancerous to immortal cells, indicating a decrease in viscosity, with smaller changes in subsequent stages. MutSα produces an effective diffusion coefficient that, coupled with the free EGFP diffusion measurements, can be used to extract a pure diffusion coefficient and a pseudo-equilibrium constant K*. The MutSα nuclear K* increased sixfold in the first stage of cancer and then decreased in the more advanced stages. The ratio of nuclear to cytoplasmic K*for MutSα increased almost two orders of magnitude in going from noncancerous to immortal cells, suggesting that this quantity may be a sensitive metric for recognizing the onset of cancer.
Advances in biophysics are providing opportunities to establish connections between cell mechanics and cellular functions. Cancer research is one area where this is especially true. Recent reports indicate that cancerous and non-cancerous cells have different physical and mechanical properties arising from biochemical alterations as normal cells transform to cancerous cells [1,2]. There is, however, significant ambiguity about how these properties change. One source of ambiguity lies with the type of cell being studied. For example, human bladder cells , mammary epithelial cells [4–6], pancreatic epithelial cells [7,8], and mouse fibroblast cells  have shown increased deformability in cancerous cells compared to normal cells; in contrast human hepatocytes , myeloid and lymphoid leukemia cells , and Lewis lung carcinoma mouse cells  have shown decreased deformability with cancer. The mechanical properties of tissues may differ from the properties of individual cells, because tissue also contains the extracellular matrix. It has also been shown that the surrounding matrix can drive cell physical development . Stiffness maps from benign human breast biopsies show uniform stiffness profiles characterized by a single peak, in contrast to malignant tissues which have broader peaks resulting from tissue heterogeneity and a characteristic low-stiffness peak representative of cancer cells . Another reason for different results is the method being used to measure the mechanical properties of cells. Atomic force microscopy has been used to test local membrane properties [11,15], while micropipette aspiration [7,16] and microplate stretchers  measure global cell properties.
We performed FRAP diffusion measurements on EGFP and MutSα-EGFP in two different cellular regions: the cytoplasm and the nucleus, and on four different cells: noncancerous, immortal, tumorigenic, and metastatic, representing four different stages of neoplastic transformation. A sample image of the fluorescence of MutSα-EGFP for immortal cells is given in Fig 2.
A summary of the diffusion measurements is given in Fig 3. For each data point, between 18 and 55 measurements were taken–see Table 3 (mean = 31.4) for a total N = 493. In the case of MutSα-EGFP, we note that these values correspond to the effective diffusion coefficient. Since the free diffusion values for MutSα-EGFP in Table 7 (labeled Df) are simply scaled versions of the corresponding free EGFP D values, there is no merit in plotting the MutSα-EGFP Df values. In the case of the MSH2 protein, we will discuss the K* results for different regions and cell types, and we will assume that these values represent the bound/free ratio (BFR) of the MutSα-EGFP complex in the nucleus. This is a reasonable assumption for the nucleus since the MMR proteins bind to the DNA in the nucleus, and in cases where a protein’s binding partner is relatively immobile (like DNA for MutSα) then the pseudo-equilibrium constant K* can be physically interpreted as the bound/free ratio of MutSα (see the discussion in Sprague, et al. ). Since the assumption that K* is the same as the BFR requires the binding partner to be nearly fixed, we will not make this assumption in the cytoplasm, where we do not know the binding partner. Several interesting observations can be made from these data. We summarize the main results in two separate tables (Tables 9 & 10), along with a brief ‘significance and context’ statement. Additional details are given in the rest of the section that follows the two tables.
In this study we have used FRAP to measure the diffusion of free EGFP and of MSH2-EGFP, which we have assumed is bound to MSH6 to form MutSα in both the cytoplasm and nucleus of four different cell types. The cells were chosen to represent, to some degree, the different stages of neoplastic transformation, namely noncancerous, immortal, tumorigenic, and metastatic.