Search

Peptide fouling is a technological drawback in filtration membrane processes. The impact of membrane characteristics on fouling by peptides from a complex whey protein hydrolysate (WPH) fouling was assessed based on advanced statistical redundancy analysis. Six membranes were characterized and tested: PES, PVDF, CF55, S11, S11+ and S11-. Among the eight physicochemical characteristics analyzed, zeta-potential (ZP) and roughness (Rz) were highly correlated with total fouling quantity (TFQ), suggesting that peptide fouling was mainly due to electrostatic interactions over wider surfaces. Concerning peptide sequences, redundancy analysis indicated that at least one characteristic among ZP, Rz, contact angle and thickness contributed significantly to the fouling of ALMPHIR, LIVTQTMK, TKIPAVFK, VLVLDTDYK, TPEVDDEALEK, TPEVDDEALEFDK or SLAMAASDISLLDAQSAPLR. It appeared that membranes with hydrophilic surfaces were more likely to be fouled by WPH regardless the peptide hydrophilicity. Besides, wider surface would enable more contact area for peptides, increasing consequently fouling. The TFQ was described by a statistical predictive model using the combined effect of ZP and Rz, of the form: ln(TFQ) = 8.64 + 0.072.XRz + 0.088.XZP (R2 = 0.9098). Moreover, statistical models were also established for each peptide, enabling to predict their specific fouling on a variety of membranes with different physicochemical characteristics.

Peptide adsorption occurring on conventional anion- and cation-exchange membranes is one of the main technological locks in electrodialysis (ED) for hydrolysate demineralization. Hence, the peptide fouling of monovalent anion (MAP) and monovalent cation (MCP) permselective membranes was studied and compared to conventional membranes (AMX-SB and CMX-SB). It appeared that the main peptide sequences responsible for fouling were TPEVDDEALEKFDK, VAGTWY and VLVLDTDYK for both anionic membranes; and ALPMHIR and TKIPAVFK for both cationic membranes. However based on the MS-MS results, the fouling was about 97–100% lower for MAP than AMX-SB and 95–100% lower for MCP than CMX-SB. This was explained by the differences in charge sign distribution at the membrane surface. Consequently, monovalent membranes can represent a very interesting opportunity for treatment of hydrolysate solution in electrodialytic processes by practically eliminating peptide fouling. At our knowledge, it was the first time that such a demonstration was done.

Peptide fouling occurring on anion- (AEMs) and cation-exchange membranes (CEMs) is one of the most serious issues of conventional electrodialysis (ED) process for hydrolysate demineralization. Nonetheless, recent studies discussed the advantages of non-conventional ED phenomena such as water splitting and electroconvection on decreasing scaling and fouling. Thereby, peptide fouling was characterized using four different ED regimes: no current applied, underlimiting (conventional), limiting (water splitting) and overlimiting (electroconvection and water splitting) conditions. Results demonstrated that fouling-related interactions were mainly electrostatic with AEMs whereas they were both electrostatic and hydrophobic with CEMs. After 60 min, the demineralization rate was six times higher in overlimiting than underlimiting conditions. In addition, peptide fouling was 62% and 36% lower in overlimiting condition for AEMs and CEMs, respectively. It was hypothesized that (1) water splitting would have repealed the peptide charges through its "barrier effect" and (2) electroconvective vortices generated at the membranes interfaces would have washed-out their surfaces and hampered the attachment of peptides. Interestingly, ED under overlimiting conditions is a promising way to avoid peptide fouling. Consequently, membranes lifetime would be longer and new ED applications would be possible.

Peptide adsorption on ion-exchange membranes (IEM) affects negatively the electrodialysis efficiency. Hence, the peptide species and nature of peptide/anion-exchange membranes (AEMs) interactions involved in fouling were explored at different pH. It appeared that fouling was severe at pH10, twice lower at pH6 and absent at pH2. Based on HPLC-MS results, the three main fouling peptides were VLVLDTDYK, TPEVDDEALEK and IDALNENK representing 86% of the total fouling. Since these peptides carried negative net charges on their C-terminal, D and E residues at pH6 and 10, interactions are of electrostatic nature and mostly happened between the negative peptide charges (COO−) and the positive charges of AEM groups (N+(CH3)3). In addition, TPEVDDEALEK interacted more intensively than VLVLDTDYK at pH6. This was explained by charges distribution at the surface of the macromolecular structures. Interestingly, working at acidic pH would completely prevent the fouling on AEMs due to repulsion between peptides and AEMs groups. To the best of our knowledge, it is the first time that an in-depth analysis with peptides from a complex protein hydrolysate was carried out to demonstrate the impact of peptide physicochemical and structural characteristics as well as the mechanisms involved in anion-exchange membranes fouling.

Peptide fouling of fifteen membranes with different physicochemical characteristics was estimated based on statistical redundancy analysis. The zeta-potential (ZP) and the roughness (Rz) parameters were highly correlated with the total fouling quantity (TFQ, RDA explanation of 86.6%), meaning that fouling by peptides was mainly due to electrostatic interactions over rough surfaces. TFQ was predicted by a statistical model using the combined effect of both zeta-potential and Rz membrane parameters, of the form: ln(TFQ) = 7.9546 + 0.1008⋅XRz + 0.03629⋅XZP (R2 = 0.8418). Concerning individual peptides and their own physicochemical properties, redundancy analysis showed that ZP, Rz, conductivity, contact angle and percentage of hydrophilic pores contributed significantly to the peptide fouling. Using the data of all 15 membranes, global predictive multivariate regression models were also established for each of the seven main peptides allowing the prediction of their specific fouling behavior on a wide range of membrane characteristics and properties. Finally, the models were validated by comparing them with experimental data of an extra membrane and consequently, peptide fouling can be well-estimated by using these relevant predictive models. Furthermore, based on the PES membranes, total fouling and individual peptide fouling were not affected by the membrane cut-off. However, based on the PVDF and PAN results, a combined effect of cut-off and material seemed to have an impact on both types of fouling: fouling decreased for these membranes when increasing their cut-off.

International audience; Polypeptide/solid charged surface interactions are omnipresent in the biomedical and biochemical fields. The present study aimed to understand the adsorption mechanisms of a cation-exchange membrane (CEM) by a well-characterized peptide mixture at three different pH values. Results demonstrated that fouling was important at pH 6, twice lower at pH 2 and negligible at pH 10. At pH 6, ALPMHIR and TKIPAVFK sequences firstly established electrostatic interactions with the negative CEM charges (SO3-) through their positive K and R residues (NH3+) creating a first nanolayer. Secondly, peptide peptide interactions occurred through their respective hydrophobic residues creating a second nanolayer. At pH 2, VLVLDTDYK and IDALNENK sequences interacted only electrostatically and that in a lower proportion since at acidic pH values, most of the CEM charges would be protonated and uncharged (HSO3) and then limit the potential electrostatic interactions. In addition, the sequences of peptides interacting at pH 2 and 6 were different. This was explained by their structure in terms of residue nature and position in the sequence. At pH 10, no fouling was observed due to the lack of positive peptide charges. To the best of our knowledge, it is the first indepth study concerning the fouling of CEMs by peptides from a complex mixture.