Tailored silica coated Ag nanoparticles for non-invasive surface enhanced Raman spectroscopy of biomolecular targets

Abstract

Silica coated Ag nanoparticles with defined surface plasmon resonances are used to selectively detect and analyze protein cofactors in solution and on interfaces via surface enhanced resonance Raman spectroscopy. The silica coating has a surprisingly small effect on optical amplification but minimizes unwanted interactions between the protein and the nanoparticle.

---

For rational design of biotechnological devices such as biosensors or biofuel cells, a profound understanding of the molecular processes between target molecules and their reaction partners is needed. Suitable in situ analytic techniques in this field should therefore be able to give structural information on the target molecules without interfering in the processes under investigation. These requirements are fulfilled by vibrational (IR and Raman) spectroscopic detection methods that have been used to study i.e. biological charge transfer reactions,^1,2 or enzymatic catalysis.^3–5 However, the complex vibrational patterns normally obtained from biological probes, make a selective detection method for the relevant processes by (i) local and (ii) target specific signal enhancement essential. The latter can be achieved by resonance Raman (RR) spectroscopy. Here the vibrational signal of a chromophore, i.e. a protein cofactor, is selectively enhanced by tuning the laser excitation line to the respective molecular electronic transition. Local enhancement on the other hand can be induced by addition of noble metal nanoparticles that are able to create surface plasmon resonances (SPRs) upon light excitation. This effect, which is especially advantageous for analysis of interfacial processes,⁶ has been largely exploited for surface enhanced Raman spectroscopy (SERS) where signal amplification of molecules in close vicinity to metal surfaces is achieved. A simultaneous local and target specific signal amplification can be obtained if the surface resonance of the nanoparticle is tuned to the molecular resonance of the target molecule. This two-fold resonance condition makes it in principle possible to follow target molecules and to analyze their interactions with other molecules or interfaces via surface enhanced resonance Raman spectroscopy (SERRS). However, intermolecular processes can easily be disturbed by the presence of the metal amplifiers; hence it has to be guaranteed that the nanoparticles, which are introduced artificially into the system, do not interfere with the molecular process under investigation. In this respect smart design of SERRS active, functional nanoparticles requires that the interactions with the target molecule are kept to a minimum, while efficient and selective enhancement is still provided. For resonance enhancement of chromophores that absorb in the blue and violet region the necessary surface enhancement can only be provided by silver nanoparticles (AgNPs). We have shown in our previous work that the SPR of AgNPs can be precisely tuned to match the molecular resonance of the model heme protein cytochrome c (Cyt c).⁷Protein denaturation was prevented by replacing the initially surface bound citrate ions with a self assembled monolayer of 11-mercaptoundecanoic acid (MUA). Cyt c was strongly bound to these Ag@MUA NPs, which resulted in a high Raman signal intensity but made it impossible to further study the interactions of the protein with other reaction partners. Silica coated metal particles have shown improve the stability of colloidal suspensions significantly^8,9 and thus could be a good alternative to monolayer protected particles. Even though the optical properties of such metal@SiO₂ particles have been studied in great detail^10,11 not much is known about their interactions with biological probes. In this work we have therefore synthesized plasmon tuned, silica coated Ag@SiO₂ NPs and studied their interaction with heme proteins. The NPs show excellent applicability in bioanalysis, as strong SERR signal enhancement of the heme cofactor is achieved, while binding between the protein and the particle is extremely weak.

For fabrication of Ag@SiO₂ NPs with defined optical properties firstly small Ag@citrate nanoparticles (d = 12 nm, c ∼ 10 nM) were synthesized by reduction of AgNO₃ with NaBH₄ in the presence of citrate as described in reference 7. A portion of these particles were subsequently used as seeds for the synthesis of larger AgNPs by adding a growth solution that contained AgNO₃ and ascorbic acid (AA) with defined concentration. Three more Ag@citrate NP batches were prepared this way by mixing 50 mL of the AgNPs seed solution with 0.2 (0.5, 1.0) mL of 1 mM AgNO₃ and 0.2, 0.5 or 1.0 mL of 1 mM AA respectively. The resulting 4 Ag@citrate NPs batches were subsequently coated with a thin silica film following the procedures given in references 12 and 13. Briefly, a freshly prepared aqueous solution of 3-aminopropyltriethoxysilane (APTS 0.4 mL, 0.2%) was added to 25 mL of seeded AgNPs at pH 5.0. Defect free coating of the SiO₂ layer was achieved by adding 2 mL of 2% sodium silicate solution. The pH was adjusted to 10.0 and the solution was stirred for 24 h. Finally, the silica coated Ag@SiO₂ NPs were purified from other reagents by centrifugation and redispersed in water before being stored at 4 °C. The Ag@SiO₂ NPs showed extremely high stability. Over the time scale of 6 months no aggregation was observed.

The particle shape was approximated by an ellipsoid exhibiting a longer (d_long) and a shorter (d_short) axis. We have shown previously that both diameters as well as their aspect ratio (d_long/d_short) are increased upon increasing the growth solution concentration.⁷ It is known that these two effects lead to a controlled red shift of the SPR position.^14,15 A further red shift of 8–10 nm was introduced by coating with SiO₂. A transmission electron micrograph (FEI Tecnai G² 20 S-TWIN, 200 kV) of Ag@SiO₂ NPs with an SPR maximum at 413 nm is shown in Fig. 1 A. The shorter diameter d_short and the mean aspect ratio of the NPs were determined to be 18 ± 2 nm and 1.2 ± 0.3 respectively (Fig. S1†), which is in very good agreement with our previously published data.⁷ From Fig. 1 A, the successful SiO₂ coating of the Ag@SiO₂ NPs is visible. However, the SiO₂ coating could be better evidenced by energy loss imaging (Zeiss LIBRA 200 FE, 200 kV), as shown in Fig. 1 B, where an average thickness of 4 ± 2 nm of the SiO₂ layer could be determined (See also Fig. S1†). In Fig. 1 C UV-vis spectra for the 4 NP batches are shown. Depending on the growth solution concentration the SPR is shifted for each batch with respective maxima at 400, 408, 413 and 421 nm. The batches are in the following named according to their SPR maximum Ag₄₀₀@SiO₂, Ag₄₀₈@SiO₂, Ag₄₁₃@SiO₂ and Ag₄₂₁@SiO₂ NPs.

Fig. 1 (A) EFTEM zero loss image and (B) EFTEM energy loss image of Ag₄₁₃@SiO₂ NPs. (C) UV-vis spectrum of the Ag₄₀₀@SiO₂ (blue), Ag₄₀₈@SiO₂ (violet), Ag₄₁₃@SiO₂ (red) and Ag₄₂₁@SiO₂ (green) NP batches. (D) Cartoon of a Ag@SiO₂ NP with Cyt c attached. The optical properties of Ag@SiO₂ NPs were optimised for enhancement of heme cofactors.

Cyt c is a small soluble redox protein that binds electrostatically to negatively charged surfaces via its lysine rich binding patch. It is therefore expected to interact also with the OH groups that exist at the SiO₂ surface. Cyt c has seen detailed characterisation in the past with SERRS^16,17 and is therefore also used in this work as a Raman marker and model protein. Cyt c was added to each Ag@SiO₂ NP solution and the respective SERR spectra were measured in a rotating cuvette with the 413 nm excitation line of a Kr⁺ laser using a confocal Raman microscope (LabRam HR-800, Jobin Yvon). The vibrational spectrum of ferric Cyt c displays a prominent band (ν₄) at Δν = 1370 cm⁻¹, which is shown in Fig. 2 A and which was used in the following to determine the Raman enhancement factor (REF).

Fig. 2 (A) RR (black) and SERR spectra (Ag₄₀₀@SiO₂ (blue), Ag₄₀₈@SiO₂ (purple), Ag₄₁₃@SiO₂ (red) and Ag₄₂₁@SiO₂ (green)) of Cyt c in the ν₄ region. (B) Raman enhancement factors (REF) of the different Ag@SiO₂ NP batches as a function of SPR maximum. The dashed lines indicate the wavelength of the excitation (λ_exc) and the Raman scattered (λ_Ra,ν4) line.

The REF can be calculated from measured RR and SERR intensities according to:

For c_RR the Cyt c concentration in solution c₀ = 20 μM was taken. The factor I_SERRS/I_RR = 6 was determined from the difference in Raman intensity with and without NPs present in solution. k is a shielding factor determined in previous work to be 0.25.¹⁸ The concentration of surface bound Cyt c (c_SERRS) in the saturation limit was measured to be 1.2 μM for the Ag₄₁₃@SiO₂ NP batch following a procedure described in reference 7 (Fig. S2†). This corresponds to a Cyt c surface concentration of 9.5 pM cm⁻² or 60% of the geometrical possible coverage.

If we assume that c_SERRS is equal for all Ag@SiO₂ NPs, REF can be derived for all NP batches from eqn (1) and is plotted in Fig. 2 B as a function of the respective SPR maximum.

REF can be seen as the product of the wavelength dependent electric field enhancement intensity EF(λ) at the exciting (λ_exc) and Raman scattered (λ_Ra) wavelengths:

REF = EF(λ_exc)EF(λ_Ra) (2)

For λ_exc = 413 nm the position of the ν₄ band corresponds to λ_Ra = 438 nm. If we assume that EF is proportional to the absorbance of the respective NP ensemble at a given wavelength, one would expect to have the highest SERR intensity for NPs that have their SPR maximum in the wavelength interval λ_exc–λ_Ra¹⁹ This is indeed confirmed in Fig. 2 B as REF clearly drops for NPs with an SPR lower than 413 nm.

The measured values for REF correspond to enhancement factors at a distance from the Ag surface that is controlled by the thickness of the SiO₂ coating. For spherical NPs in a homogeneous medium the distance dependence of REF is given by:²⁰

where a₀ is the radius of the nanoparticle, d is the distance of the Raman scatterer to the metal surface and REF(0) refers to the Raman enhancement factor for the probe molecule in direct contact with the surface. REF(0) was theoretically determined to be ca. 1000.²¹ Applying eqn (3) for a₀ = 9 nm and a distribution of d(SiO₂) taken from Fig. S1 E† one would expect the average value of REF to decrease down to 20. Surprisingly, the measured value of REF for Ag₄₁₃@SiO₂ NPs is 20 times higher than theoretically predicted. This effect may partially be explained by the different dielectric constants of the SiO₂ coating and its aqueous environment, which could lead to an increase in the field enhancement at the SiO₂/H₂O interface as predicted theoretically.^22,23 Surprisingly, REF is also 3 times higher than measured for Ag@MUA NPs.⁷ The latter exhibited a similar Cyt c surface coverage and an even smaller layer thickness of 2 nm.

It should be noted that the SPR could also be tuned to 413 nm by using smaller Ag NPs and thicker SiO₂ coatings (see Fig. S1 B and S3†) but in these cases REF was distinctly smaller. For example, REF of Ag₄₁₃@SiO₂ NPs with d_short = 17 ± 5 nm and d(SiO₂) = 6 ± 2 nm was determined to be ca. 100. For Ag₄₁₃@SiO₂ NPs with 15 nm SiO₂ coatings no SERR enhancement could be observed any more (Fig. S3†).

For applications of Ag@SiO₂ NPs as non-invasive amplifiers it is required that binding of target proteins to the NPs is weak enough not to interfere with the processes under investigation. The binding properties of Cyt c to the Ag₄₁₃@SiO₂ particles were therefore tested by UV-vis spectroscopy upon adding and removing the NPs in a Cyt c containing solution as shown in Fig. 3. The data were compared with similar measurements using Ag₄₁₃@MUA NPs from ref. 7.

Fig. 3 Black lines: Uv-vis spectra of (A) Ag₄₁₃@SiO₂ NPs in solution, (B) Cyt c, (C) Ag₄₁₃@SiO₂ NPs together with Cyt c and (D) after centrifugation of the NPs. The green lines correspond to identical measurements using Ag₄₁₃@MUA NPs (data taken from ref. 7).

In Fig. 3 A and B the UV-vis spectra of the NPs (Ag₄₁₃@SiO₂ (black) and Ag₄₁₃@MUA (green)) and Cyt c in separate solutions are shown respectively. In both cases the characteristic absorption peaks around 413 nm for the NPs and 410 nm for ferric Cyt c can be seen. If both, Cyt c and NP, solutions are put together a red shift in the SPR for Ag₄₁₃@MUA NPs is observed but not for Ag₄₁₃@SiO₂ NPs (Fig. 3 C). In the case of Ag₄₁₃@MUA the SPR shift is most likely caused by a change in the local dielectric environment of the particle, indicating the formation of a strong Ag@MUA-Cyt c complex. The UV-vis spectrum of Ag@SiO₂ NPs with Cyt c on the other hand is a superposition of the two isolated systems shown in Fig. 3 A and B, leading to the conclusion that no strong interacting NP–protein complex is formed. This assumption is further confirmed by Fig. 3 D, which shows the UV-vis spectra of the solutions after centrifugation of the NPs. In the case of Ag₄₁₃@SiO₂ NPs Cyt c remained in solution, while it was centrifuged away with the NPs in the case of Ag413@MUA NPs. Based on these observations we conclude that the interaction of Ag₄₁₃@SiO₂ NPs with Cyt c is much weaker than for Ag₄₁₃@MUA NPs.

The biocompatibility of the Ag@SiO₂ NPs was tested by comparing the SERR spectrum of Cyt c with its RR spectrum in solution (Fig. 4 A and B). The high similarity between the two spectra indicates that no alteration of the heme pocket structure occurred upon surface attachment. This observation can also be seen as an additional proof for the weak binding between Cyt c and the NPs. Surprisingly strong alterations in the spectral pattern were observed when the same SiO₂ coating procedure was applied to rough Ag electrodes. In this case a significant contribution from a non-native high spin species²⁴ was observed, indicated by an additional band at 1490 cm⁻¹ (Fig. 4 C). This denaturation effect has already been reported in a previous work using a slightly different SiO₂ coating procedure²⁵ and is attributed to the fact that on rough metal surfaces a non-complete SiO₂ layer is formed and thus spectral contributions of Cyt c bound directly to the Ag surface are observed. Since these proteins are very close to the Ag surface even a small amount of this fraction can dominate the SERR spectrum.

Fig. 4 (A) RR spectrum of Cyt c in solution. (B, C) SERR spectra of Cyt c bound to Ag₄₁₃@SiO₂ NPs and on rough Ag@SiO₂ electrodes. (D) SERR spectrum of YCyt c covalently bound to flat Au electrodes after Ag@SiO₂ NP attachment.

The Ag@SiO₂ NPs were finally used to enhance the signal of proteins bound to plasmon inactive surfaces. As a reference system yeast cytochrome c (YCyt c) immobilised on a flat Au surface was chosen. Covalent binding of YCyt c to the bare Au electrode was achieved via its single surface cysteine Cys102 as follows: a cleaned Au electrode was immersed into 35 μM aqueous solution of YCyt c for 1 h followed by immersion into 2 M KCl solution for 20 min to remove the electrostatically bound proteins and rinsing with water.

Subsequently performed SERR experiments of the Au-YCytcelectrode did not give any Raman signal. After adding 20 μL of Ag@SiO₂ NP solution the signal increased significantly with time, which we attribute to a slow deposition of the NPs on the Au surface. After approximately 1 h a stable high quality Raman spectrum could be obtained (Fig. 4 D). The signal exclusively arises from proteins bound to the Au surface as the weak attraction of the NPs is not able to break the strong thiol bond between YCyt c and Au. The spectrum of the covalently bound YCyt c showed a significant contribution from the non-native high spin species, which is in agreement with previous SERR measurements of YCyt c bound viaCys102 to Au–Ag hybrid electrodes.²⁶

In summary we have shown that Ag@SiO₂ NPs with tunable optical properties are able to selectively enhance protein chromophores while, due to the very weak interaction with the target molecule, the properties of the protein are not affected. In comparison to monolayer protected Ag@MUA NPs the silica coating leads to higher SERR intensity upon a much lower protein binding affinity. These NPs are therefore very promising candidates as non-invasive optical amplifiers for the spectroscopic analysis of biomolecular processes. Preliminary measurements also show that these NPs can be used for amplification of target molecules bound to plasmon inactive surfaces.

Acknowledgements

The authors would like to thank Peter Hildebrandt for helpful discussions. Also the ZELMI (TU Berlin) and the Helmholtz-Zentrum Berlin are gratefully acknowledged for access to the TEM facilities and particularly Soeren Selve and Markus Wollgarten for helpful discussions. Financial support from the Fonds der Chemie (IW), the Alfried Krupp Wissenschaftskolleg, Greifswald (AS) and the DFG (Unicat) is gratefully acknowledged.

References

1. H. K. Ly, M. Sezer, N. Wisitruangsakul, J. J. Feng, A. Kranich, D. Millo, I. M. Weidinger, I. Zebger, D. H. Murgida and P. Hildebrandt, FEBS J., 2011, 278, 1382.
2. C. Kotting and K. Gerwert, ChemPhysChem, 2005, 6, 881.
3. D. Millo, M. E. Pandelia, W. Lubitz, P. Hildebrandt and I. Zebger, Angew. Chem., Int. Ed., 2011, 50, 2632.
4. T. Ogura and T. Kitagawa, Biochim. Biophys. Acta, Bioenerg., 2004, 1655, 290.
5. R. M. Nyquist, D. Heitbrink, C. Bolwien, R. B. Gennis and J. Heberle, Proc. Natl. Acad. Sci. U. S. A., 2003, 100, 8715.
6. J. F. Li, Y. F. Huang, Y. Ding, Z. L. Yang, S. B. Li, X. S. Zhou, F. R. Fan, W. Zhang, Z. Y. Zhou, D. Y. Wu, B. Ren, Z. L. Wang and Z. Q. Tian, Nature, 2010, 464, 392.
7. A. Sivanesan, H. K. Ly, J. Kozuch, M. Sezer, U. Kuhlmann, A. Fischer and I. M. Weidinger, Chem. Commun., 2011, 47, 3553.
8. R. A. Caruso and M. Antonietti, Chem. Mater., 2001, 13, 3272.
9. P. Mulvaney, L. M. Liz-Marzan, M. Giersig and T. Ung, J. Mater. Chem., 2000, 10, 1259.
10. V. Salgueirino-Maceira, F. Caruso and L. M. Liz-Marzan, J. Phys. Chem. B, 2003, 107, 10990.
11. T. Ung, L. M. Liz-Marzan and P. Mulvaney, J. Phys. Chem. B, 2001, 105, 3441.
12. L. M. LizMarzan, M. Giersig and P. Mulvaney, Langmuir, 1996, 12, 4329.
13. T. Ung, L. M. Liz-Marzan and P. Mulvaney, Langmuir, 1998, 14, 3740.
14. S. Link and M. A. El-Sayed, J. Phys. Chem. B, 1999, 103, 8410.
15. K. L. Kelly, E. Coronado, L. L. Zhao and G. C. Schatz, J. Phys. Chem. B, 2003, 107, 668.
16. D. H. Murgida and P. Hildebrandt, Acc. Chem. Res., 2004, 37, 854.
17. J. J. Feng, D. H. Murgida, U. Kuhlmann, T. Utesch, M. A. Mroginski, P. Hildebrandt and I. M. Weidinger, J. Phys. Chem. B, 2008, 112, 15202.
18. P. Hildebrandt and M. Stockburger, J. Phys. Chem., 1986, 90, 6017.
19. A. D. McFarland, M. A. Young, J. A. Dieringer and R. P. Van Duyne, J. Phys. Chem. B, 2005, 109, 11279.
20. F. Siebert and P. Hildebrandt, Vibrational Spectroscopy in Life Science, Wiley-VCH, Weinheim, 2008.
21. E. J. Zeman and G. C. Schatz, J. Phys. Chem., 1987, 91, 634.
22. J. J. Feng, U. Gernert, M. Sezer, U. Kuhlmann, D. H. Murgida, C. David, M. Richter, A. Knorr, P. Hildebrandt and I. M. Weidinger, Nano Lett., 2009, 9, 298.
23. C. David, M. Richter, A. Knorr, I. M. Weidinger and P. Hildebrandt, J. Chem. Phys., 2010, 132, 024712.
24. S. Oellerich, H. Wackerbarth and P. Hildebrandt, J. Phys. Chem. B, 2002, 106, 6566.
25. J. J. Feng, U. Gernert, P. Hildebrandt and I. M. Weidinger, Adv. Funct. Mater., 2010, 20, 1954.
26. M. Sezer, J. J. Feng, H. K. Ly, Y. F. Shen, T. Nakanishi, U. Kuhlmann, P. Hildebrandt, H. Mohwald and I. M. Weidinger, Phys. Chem. Chem. Phys., 2010, 12, 9822.

---

Footnote

† Electronic supplementary information (ESI) available: Further characteristics of Ag@SiO₂ NPs from TEM measurements. SERR intensities of Cyt c as a function of concentration. See DOI: 10.1039/c1ra00781e

---