Autor(es):
Morão, A. 
; Nunes, J. C. ; Sousa, Fani ; Amorim, M. T. Pessoa de ; Escobar, I. C. ; Queiroz, J. A.
Data: 2011
Identificador Persistente: http://hdl.handle.net/1822/15099
Origem: RepositóriUM - Universidade do Minho
Assunto(s): Ultrafiltration; Plasmid DNA; Monte Carlo simulations; Sieving coefficients; Modeling
Descrição
A mass transfer model is proposed for predicting sieving coefficients, Sobs, of supercoiled plasmid DNA (pDNA), in the presence of a salt, in membranes with narrow pores, i.e., pores smaller than the gyration radii of the plasmids to be considered for a certain application. The model assumes that permeation occurs due to plasmid suction at the membrane surface as a result of the convective flow, being the probability of permeation also dependent on the instantaneous molecular conformation of the plasmid, when getting close to the pore. Two different approaches are tested to model plasmid structure, that of a closed segmented chains (CSC) of double stranded DNA, and that of considering the superhelical chain as a freely jointed chains (FJC). Both approaches were used to estimate the radius of gyration, rg, of different plasmids by statistical simulation, and the obtained values were compared with experimental data available in the literature.
A 6050 bp plasmid, pVAX1-LacZ, was used in the experimental work, in which filtration tests were performed using three different ultrafiltration membranes of known pore size, in a 10 ml stirred cell. At constant ionic strength, sieving coefficients were determined as a function of the permeate flux, Jv, at two different values of stirring speed, ω. The results are in very good agreement with the model predictions at the highest stirring speed and the observed deviations found at the lowest stirring speed were interpreted with the aid of the developed model by considering the possibility of plasmid adsorption. Then, it was investigated the effect of changing the ionic strength of the medium at constant Jv and ω. The obtained results clearly agree with the model predictions.
