Our major research goal this summer is to find out what conditions are necessary for reliable SERS of proteins. Lots of people are interested in this for many different reasons. Ours are to further collaborations with some colleagues in the biology department who are interested in finding out some structural information about proteins they are studying. Hopefully as we work on this we’ll be able to get some general information about what one needs to think about in doing SERS with proteins.

We’ve started by trying to reproduce literature results. You’d think that would be easy, but it isn’t turning out to be the case. As an example, consider lysozyme. I have four published reports of lysozyme SERS, and all four have markedly different SERS spectra. Each is reviewed in chronological order below the fold.

Paper 1: Hu, Sheng, Xu and Zeng; Surface enhanced Raman spectroscopy of lysozyme, Spectrochimica Acta, 51A 1087-1096, 1995.

Experimental Details
Hu et al. used lysozyme from egg white. Their colloids were reduced with sodium borohydride, and aged for several weeks before use. Lysozyme solutions were prepared by dissolving the solid in triply distilled water, and pH was adjusted by adding HNO₃, HCl or NaOH. Their Raman excitation source was an Ar⁺ laser at 514.5 nm.

Results
No normal Raman spectra of lysozyme are presented. Several SERS spectra are though, over a range of pH values (lyzosyme’s pI is reportedly between 9 and 11). The strongest SERS signals were observed when the pH was between 2 and 5. From their spectra, they infer (via the frequencies of the Amide I and III bands) that lysozyme adsorbs to the colloids via the α helix, and that most of the aromatics in the side chains are oriented perpendicular to the surface (they assigned several strong bands to the Phe, Trp, Tyr and His residues).

Paper 2: Stewart and Fredericks; Surface enhanced Raman spectroscopy of peptides and protiens adsorbed on an electrochemically prepared silver surface., Spectrochimica Acta A, 55 1615-1640, 1999.

Experimental Details
Lysozyme was purchased from Sigma, used as received by dissolving in triply distilled water. SERS surfaces were prepared electrochemically (method described in a reference). Raman excitation via a He-Ne laser at 632.8 nm.

Results
The lysozyme SERS spectrum is dominated by an intense band at 1641 cm^-1 assigned as an Amide I vibration. The majority of the other bands are from aromatic amino acid residues (Phe, Trp, Tyr and His).

Paper 3: Podstawka, Ozaki and Proniewicz; Adsorption of S-S Containing Proteins on a Colloidal Silver Surface Studied by Surface-Enhanced Raman Spectroscopy, Applied Spectroscopy, 58 1147-1156, 2004.

Experimental Details
Egg white lysozyme was purchased from Wako Chemical Co. Solutions were prepared from the samples as received with deionized water. Silver colloids were made by borohydride reduction (both using the procedure described by Hu et al. and Lee and Meisel). Raman excitation was an Ar⁺ laser at 514.5 nm.

Results
They report a normal Raman and SERS spectra of lysozyme (normal Raman is of the solid). Their SERS spectrum matches the normal Raman spectrum fairly well, but is markedly different from that reported by Hu et al., and somewhat different from that of Stewart and Fredricks [1]. They observe a relatively strong S-S vibration, and as others have mentioned, peaks due to Trp, Phe, His, Tyr, and Amide I and III.

Paper 4: Han, Huang, Zhao and Ozaki [2]; Label-Free Highly Sensitive Detection of Proteins in Aqueous Solutions Using Surface-Enhanced Raman Scattering, Analytical Chemistry, 81 3329-3333, 2009.

Experimental Details
Egg white lysozyme was purchased from Wako Chemical Co. Solutions were prepared in a phosphate buffer (NaH₂PO₄:Na₂HPO₄, pH = 7.0). Silver colloids were made by citrate reduction. Prior to mixing with colloids, the lysozyme was mixed with a 0.1 M sodium sulfate solution (to aggregate the colloids), and the pH was adjusted to 3.0. A 785 nm NIR laser was used for Raman excitation.

Results
A set of three SERS spectra are reported, as a function of increasing lysozyme concentration, along with a solid state spectrum. The spectra presented here do not resemble any of the spectra in the papers described above. For the SERS spectra, this might not be surprising (although it is somewhat troubling). That the normal Raman spectrum of the solid is not at all like the spectra of solid lysozyme published elsewhere is somewhat confusing. There is little to no evidence of the Amide I or III vibrations, although the authors assign some peaks that are barely out of the noise to them.

Overall
So what does all of this mean? In one sense, it looks like obtaining consistent SERS spectra of proteins might be a pipe dream. Four papers, four different SERS spectra of just one protein (and a fairly small one at that). This really illustrates some of the problems with SERS (that are all well known). Results depend on the Ag substrate, and it can be difficult to consistently prepare colloids that are “the same” batch to batch. Some of the differences in the results of these four papers may be due to that. Some may be due to instrumental response, and different excitation wavelengths (I don’t think this is too likely).

[1] This may be due to different laser sources, and coupling to the local surface plasmon. There are also significant differences between the types of surfaces used in these three studies.

[2] The same Ozaki as in the previous paper.