Determinación del precio FOB

Biomineralisation of silica in diatoms occurs by a class of homologically-conserved proteins, silaffins, that have been extensively studied [Kroger et al. [1999, 2001, 2002]; Sumper and Brunner [2008]]. Silaffin mimics, the most studied of which is the R5 peptide [Kroger et al. [1999]], have been widely used by several groups to control the synthesis of silica nanostructures [Brott et al. [2001]; Patwardhan and Clarson [2002]; Lutz et al. [2005]; Soto-Cantu et al. [2010]; Xia and Li [2011]; Pat- wardhan [2011]]. In addition, peptides capable of binding to silica nanoparticles (either those synthesized biomimetically [Naik et al. [2002a]] or using standard pro- cedures [Patwardhan et al. [2012]; Puddu and Perry [2012]]), the (100) α-quartz surface [Tamerler et al. [2007]] and thermally grown silica films [Eteshola et al. [2005]] have been identified using biocombinatorial techniques (Table 1.1).

Positively charged residues are prevalent in many of the silica binding pep- tide sequences reported [Brott et al. [2001]; Patwardhan and Clarson [2002]; Naik et al. [2002a]; Eteshola et al. [2005]; Sano and Shiba [2003]; Xia and Li [2011]; Patwardhan et al. [2012]; Puddu and Perry [2012]]. For example, silaffins are en- riched in lysine residues [Kroger et al. [1999, 2001]]. However, in silaffins lysines are post-translationally alkylated, and so their interactions with silica could be hy- drophobic in character rather than mediated by the terminal ammonium group.

In addition, it has been postulated that the zwitterionic nature of silaffins, with positively-charged ammonium groups and negatively-charged phosphorylated ser- ines, aids silaffin aggregation, rather than the charged groups necessarily having a direct influence on silica precipitation [Kroger et al. [2002]; Sumper and Brunner [2008]]. Early experimental evidence to support this hypothesis was the observation that silaffins formed aggregates of approximately 700 molecules at low ionic strength [Kroger et al. [2002]]. Later studies have shown that phosphate or sulfate ions can induce aggregation of synthetic polyamines in a concentration and pH dependent manner [Brunner et al. [2004]; Lutz et al. [2005]]. The amount and size of the silica nanoparticles precipitated from silicic acid solution by these polyamines correlated with polyamine aggregate size. Hence the prevalence of positively charged residues in silica precipitating peptides reported in the literature [Brott et al. [2001]; Pat- wardhan and Clarson [2002]; Sano et al. [2005]; Hayashi et al. [2009]; Xia and Li [2011]] is not necessarily an indication that they interact directly with silica surfaces because in each case a phosphate-containing buffer was used.

Moreover, in the case of biocombinatorally selected silica-binding peptides, the nature of the positively charged residue enriched differs depending on the se- lection conditions. Hayashiet al. used AFM to measure adsorption forces between ferritin-fused mutated versions of the N-terminal hexapeptide of TBP-1 (RKLPDA) and silica surfaces [Hayashi et al. [2009]]. In otherwise identical conditions, the R1K mutant (the TBP-1 hexapeptide in which the first residue, arginine, was replaced by lysine) adsorbed less strongly to silica than the wild-type. In the pioneering work by Mirau et al. however, peaks in the STD NMR spectrum of TBP-1, the full 12 residue peptide, indicate that hydrogens at the charged ends of both the R1 and K2 side-chains, as well as those belonging to the methyl groups of L3 and A6, were in close proximity to the surface of a silica nanoparticle when adsorbed [Mirau et al. [2011]]. Histidine, rather than lysine or arginine, was enriched in peptides identified by phage display experiments carried out by Eteshola et al. [Eteshola et al. [2005]]. Under the selection conditions employed in the latter study, pH 7.4, histidine is predominantly present in solution in its neutral, rather than protonated, form. These findings, taken together with the prevalence of hydrophobic residues in some sequences, especially those that bind strongly to quartz (16-20 Table 1.1), suggest that at least two different peptide-silica binding mechanisms could exist– one mediated by electrostatics and the other by hydrophobic shielding; a hypothesis supported by the recent work of Puddu and Perry [Puddu and Perry [2012]].

To probe the intrinsic affinity of individual residues within a peptide for silica, independently of peptide conformational effects, experiments to study amino-

sequence substrate

1 R5 SSKKSGSYSGSKGSKRRIL1 _{PD: silafin derivative}

2 MSPHPHPRHHHT2 PD: silica ppt. by R5 3 RGRRRRLSCRLL2 PD: silica ppt. by R5 4 HPPMNASHPHMH3 _{PD: thermally grown SiO}

2

5 HTKHSHTSPPPL3 PD: thermally grown SiO2

7 TBP-1 RKLPDAPGMHTW4 PD: NP (surface area 0.4 m2kg−3₎

8 HKKPSKS5 _{PD: SiO}

2 and TiO2 NPs (∼50 nm)

9 TKRNNKR5 PD: SiO2 and TiO2 NPs (∼50 nm)

10 YITPYAHLRGGN6 PD: NP (15.0_±0.4 nm) 11 KSLSRHDHIHHH6 _{PD: NP (82.1±3.6 nm)} 12 LDHSLHS6,7 PD: NP (82.1_±3.6 nm) 13 MHRSDLMSAAVR6 PD: NP (450_±22 nm) 14 KLPGWSG7 _{PD: NP (82±4 nm)} 15 AFILPTG7 PD: NP (82_±4 nm) 16 RLNPPSQMDPPF8 PD: (100) quartz 17 QTWPPPLWFSTS8 _{PD: (100) quartz}

18 QBP-1 PPPWLPYMPPWS9 Designed: Quartz 19 LPDWWPPPQLYH9 Designed: Quartz 20 SPPRLLPWLRMP9 _{Designed: Quartz}

Table 1.1: Silica-binding peptide sequences along with either the substrate against- or method by which they were selected (‘PD’ denotes phage display and ‘NP’ denotes nanoparticle). Positively charged residues have been highlighted in red, negatively charged ones in blue and polar in green. Aromatic residues have been underlined. 1 Reference Kroger et al. [1999],2 Reference Naik et al. [2002a],3 Reference Eteshola et al. [2005],4_{Reference Sano et al. [2005],}5_{Reference Chen et al. [2006],}6_Reference

Patwardhan et al. [2012],7 Reference Puddu and Perry [2012],8 Reference Tamerler et al. [2007],9 Reference Oren et al. [2007].

acid adsorption onto various silica surfaces have been conducted by several groups [Zimmerman et al. [2004]; Churchill et al. [2004]; Vlasova and Golovkova [2004]; Alaeddine and Nygren [1996]; Kitadai et al. [2009]; Lopes et al. [2009]]. Churchillet al. found that lysine bound to quartz more strongly than the other amino-acids they studied (tyrosine, alanine, glycine, glutamate and aspartate, within the pH range 2–11) [Churchill et al. [2004]], whilst Vlasova and Golovkova found that arginine bound most strongly to highly dispersed silica surfaces out of the positively charged residues, lysine, ornithine, arginine and histidine (within the pH range 2–8) [Vlasova and Golovkova [2004]]. Contrary to Churchill et al. and Vlasova and Golovkova, however, Zimmermanet al. reported that neither lysine nor dilysine adsorbed onto mesoporous and nonporous silica, unlike the hydrophobic amino-acid tryptophan; this was found to interact strongly with both surfaces (at pH 5.7) [Zimmerman et al. [2004]]. The former result corroborates the work of Alaeddine et al. who measured the free energy of adsorption of the amino-acids leucine and serine onto both hydrophilic and hydrophobic quartz surfaces. They found that the binding affinity was greater for the former, non-polar amino-acid at both interfaces (at pH 7.2) [Alaeddine and Nygren [1996]]. In a more recent attenuated total reflectance IR spectroscopy study, 81_±5 % of the L-lysine molecules adsorbed to an amorphous sil- ica surface were in the cationic state, with the remaining 19_±5 % in the zwitterionic state (pH 7.1 - 9.8) [Kitadai et al. [2009]]. Although Kitadai et al. concluded that primarily the electrostatic attraction between lysine and the silica drives adsorption, the fact that at pH below 8.5 a higher proportion of adsorbed lysine is in the zwitte- rionic state than lysine free in solution suggests that hydrophobic interactions may also be important. Overall, the above results highlight that the exact nature of the silica surface investigated–e.g. amorphous v.s. crystalline, silanol density, silanol type (vicinal, geminal or isolated) and charge state–can have a profound effect on adsorbate affinity. In addition, it must be noted that comparison between studies at a specific silica interface (e.g. quartz) is hindered due to the impact of sample preparation on surface properties [Kosmulski [2002]].

As well as surface binding affinities, the orientation and geometry of selected amino-acids adsorbed onto silica has also been experimentally investigated. Employ- ing SFG, the adsorption of 8 amino-acids (phenylalanine, leucine, glycine, lysine, arginine, cysteine and alanine) onto a model hydrophilic SiO2 surface was probed

[Holinga et al. [2011]]. For all adsorbates, C–H vibrational modes were absent from the SFG spectra. Since QCM measurements confirmed the presence of the amino- acids at the interface, this observation suggests a lack in the net orientation and ordering of the adsorbed molecules. In dry conditions, on the other hand, alanine

and glycine were found to bind to amorphous silica in a specific geometry: the am- monium terminus was co-ordinated to 3 or 4 surface Si sites in each case, while the carboxylate terminus pointed away from the interface and exhibited slightly more orientational freedom [Ben Shir et al. [2010, 2012]]. However, under conditions more comparable to those used in the SFG experiments performed by Holingaet al. (i.e.

ambient temperature and in the presence of water) the results of both SFG and NMR experiments were consistent; the NMR spectra of alanine and glycine suggest that surface desorption is likely to occur and that the adsorbates are orientationally disordered at the interface.

Experimental [Hayashi et al. [2009]] and bioinformatics [Oren et al. [2007]] studies have shown that peptides with identical amino-acid content but different sequences have different surface binding characteristics, underlining the role of pep- tide conformation on adsorption. This is exemplified by recent MD simulations in which the calculated gold-binding affinity of several peptides for gold were not sim- ple sums of their constituent amino-acids binding energies [Verde et al. [2011]; Feng et al. [2012]]. Proline, known to restrict the conformational freedom of a peptide backbone due to its ring structure, features widely in both predominantly positively charged and charge neutral silica-binding peptides (Table 1.1). Its presence could indicate peptide stiffness is a pre-requisite to strong silica binding–the opposite of that hypothesized to be the case at the gold interface [Hnilova et al. [2008]; So et al. [2009]; Verde et al. [2011]].