*2.1. Protein*

The accurate and sensitive detection of proteins is crucial for proteomics and therapeutic research. Proteins are essential parts of organisms that are involved in virtual process with various functions. Therefore, structure analysis and quantitation of major species of functional proteins are critically important for understanding their functions. Meanwhile, there is an urgen<sup>t</sup> demand for rapid and sensitive analysis of disease biomarkers for the early diagnosis of diseases and raising the survival rate of patients.

### 2.1.1. Crucial Functional Proteins

Albumins generally act as transport proteins for numerous compounds in blood plasma [26]. Bovine serum albumin (BSA) was chosen as a model molecule to research the enhancement of SERS-active substrates in label-free detection. Noble metallic nanomaterials such as gold nanocylinders [22] and SiO2/Au nanoshells [17] were used as substrates for BSA detection. To keep proteins in their natural structure and conformation, in-liquid SERS detection of protein was based on nanoparticle aggregating as SERS-active hot spots for label-free detection [27]. Foti et al. [26] applied optically induced gold nanorods aggregation to the detection of BSA in liquid by combining light scattering, plasmon resonance and SERS. Other Raman resonant biomolecules such as catalase and hemoglobin were used to investigate the application of this methodology, and the limits of detection of catalase and hemoglobin were much lower than that of BSA.

Hemoglobins are a kind of iron-containing metalloproteins that transport oxygen in erythrocytes of all vertebrates. Iron porphyrins in hemoglobins can be an intrinsic selective probing which shows resonance enhanced Raman spectrum when excited with a laser in Soret and Q-band regions [15]. Therefore, label-free detection strategy could be applied to hemoglobin detection. Silver nanomaterials, such as colloidal AgNPs [15] and nano-silver film [28], were usually used as substrates in SERS experiment to achieve excellent enhancement of Raman signal. A novel SERS nanoprobe modified with organic cyanide (4-mercaptobenzonitrile, MBN) was developed for accurate detection of oxidized hemoglobin and Fe3+ ions in living cell [7]. The MBN-modified, SERS-active nanopipette gave new insights into in situ single cell detection.

Enzymes, which are mostly proteins in nature and act as selective biological catalysts, were characterized by SERS. Thrombin is one kind of serine proteolytic enzyme that plays an important role in regulating blood coagulation. Due to the successful selection of the thrombin-binding aptamer with high affinity and high specificity, a lot of aptamer-based SERS methods were developed for tracing thrombin analysis with improved selectivity and sensitivity [25,29]. The thrombin-binding aptamer was used as the capture probe with selective affinity that realized the SERS application in complex matrices. Meanwhile, by using the unique catalytic activity of thrombin, the enzymatic amplification-based strategy was established to achieve sensitive and selective detection of thrombin. The addition of thrombin could catalytically cleave the multiple arginine peptides into fragments, so that the aggregates of gold nanoparticles (AuNPs) incorporating Raman reporter molecules induced by peptides was decreased [30]. As the ability of fragments to form "hot spots" being weakened, the SERS signals were sharply diminished. The limit of detection (LOD) was 160 fM, which indicated that the sensitivity of this method was improved compared with non-enzymatic amplification based methods. This strategy can also be applied to other enzymes by appropriately utilizing the catalytic properties. Kinases are a group of enzymes that catalyze phosphorylation of specific substrate proteins [31–33]. Using this property, the kinase was detected by monitoring the changes in the intensity of Raman peak markers before and after phosphorylation [34]. The quantification of activity level of the enzyme can be achieved utilizing the enzyme concentration based on the enzymatic amplification-based strategy [35].

Immunoglobulin, also known as the antibody, is Y-shaped protein in different varieties known as isotypes [36,37]. Immunoglobulin is generated by the immune system to counteract pathogens such as viruses [38] and bacteria [39]. Based on the inherent antigen-antibody reaction, the immunoassay-based strategy was employed to detect immunoglobulin [40]. The traditional sandwich-format assays were usually achieved by combining antibody modified SERS-active substrates to capture target immunoglobulins and extrinsic Raman labels to enhance antigen-antibody binding kinetics [41]. Neng et al. [42] utilized AuNPs coated with the antigen as the SERS-active substrate and protein A/G modified with Raman reporter molecules as a bi-functional reporter to detect the antibody of West Nile Virus. The LOD was 2 ng/mL antibodies in the serum sample. Variants of this detection platform were applied to the trace-level detection of immunoglobulin.

Cytokines are secreted proteins produced by immune system cells regulating the immune activity [43,44]. Immunoassay-based strategy was also applicable to cytokines detection. Kami ´nska et al. [45] developed an immunoassay-based strategy employing diatom biosilica as the SERS-active substrate and DTNB-labeled immune-AuNPs as the SERS nanotag to detect the interleukin 8 (IL-8) in blood plasma [46]. Based on this method, a SERS immunoassay combined with a microfluidic device was developed for multiplexed recognition of interleukins IL-6, IL-8, and IL-18 in blood plasma with the LOD of 4.2 pg/mL. Wang et al. [47] also developed a multiplexed immunoassay of three cytokines, interferon gamma, interleukin-2, and tumor necrosis factor alpha, based on SERS signal enhanced by the controlled assembly of "hot spot" with low LOD and large signal-to-noise ratio in complex matrix.

Hormones communicate among organs and tissues to regulate various functions as a class of signaling molecules [48]. Human chorionic gonadotropin (hCG), an important pregnancy diagnostic marker, is a glycoprotein hormone that stimulates steroid hormone and progesterone production in the luteum [49]. Liang et al. [18] developed a series of methods based on the nanogold reaction between HAuCl4 and H2O2 that forms AuNPs with Victoria blue 4R molecular probes to detect hCG. The concentration of hCG can influence the catalytic effect of nanozymes such as AgNPs clusters and GO. Therefore, the changes of SERS intensity were linear to the concentration of hCG. This novel strategy can be further applied to other proteins combined with the nanozyme catalysis to develop the SERS detection platform.

Receptors are protein molecules that bound with extracellular chemical signals, causing some forms of cellular/tissue responses [50]. G-protein-coupled receptor 120 (GPR120) mediated response to long chain fatty acids (LCFAs), but the mechanism of GPR120 acting in the transduction of LCFAs was uncovered. To understand the function of GPR120 in fat chemoreception, SERS-active gold nanorods conjugated with fluorescence-active CaMoO4:Eu3+ NPs and Raman reporter molecule 4-mercaptobenzoic acid were used to achieve SERS-fluorescence imaging of GPR120 in a single cell [51].

### 2.1.2. Disease Biomarkers

Disease biomarker refers to an extracellular indicator of disease biological state or condition. These traceable substances in the organism indicate organ functions or other aspects of health. The quantitative and sensitive detection of biomarkers is very meaningful for early clinical diagnosis and evaluating the therapeutic response. Early diagnosis has grea<sup>t</sup> potential in improving the survival of patients with serious diseases, such as cancer, neurodegenerative disorders, and cardia-cerebrovascular diseases. SERS-based assays have attracted significant attention as a highly sensitive and non-destructive analysis with grea<sup>t</sup> potential in biomarker detections.

Cancer is a kind of malignant disease curable only in the early stage. Cancer biomarkers, typical group of proteins, are usually over-expressed during tumor progression [16]. Therefore, tumor biomarkers have already been used for specific cancer identification, including carcinoembryonic antigen (CEA) for colorectal cancer [24], α-fetoprotein (AFP) for liver cancer [16] and prostate specific antigen (PSA) [52]. "Sandwich" structure SERS immunoassays were usually used in biomarker protein detection and involved SERS-active substrates, target proteins, and SERS nanotags [53]. The multiple

SERS-active substrates were prepared by conjugating antibodies to capture target proteins [3,24,54]. Then the nanostructure-based SERS nanotags modified with Raman reporter and target proteins participated the antibody-antigen-antibody interaction. Li et al. [24] utilized gold nanobowl arrays and gold nanoshells as SERS-active substrates and SERS nanotags, respectively. This sandwich immunoassay showed a good linear relationship between CEA concentration and SERS signal intensity with a LOD of 1.73 pg/mL CEAs. PSA was detected on a glass slide with spot-arrays using SERS dot as the probe (Figure 2A). A full-area confocal raster Raman was applied to detect the SERS dot at the single probe level fixed on the glass slide by antigen-antibody interaction. This method showed high sensitivity with a LOD of 3.4 fM PSAs [52]. Besides single target detection, the SERS-based multiplex immunoassays were utilized to the ultrasensitive detection of different cancer biomarkers simultaneously. Wang et al. [16] used two kinds of Raman reporters labeled SiO2@Ag immune probes and gold-film hemisphere array substrate to realize multiple detection of PSA and AFP (Figure 2B). Gao et al. [55] developed a droplet-based microfluidic SERS sensor based on wash-free magnetic immunoassay technique. Using the microfluidic system integrated compartments for mixing, PSA in serum can be split and detected without any washing steps. Quantitative evaluation of PSA was realized using the Raman signals of the residual SERS nanotags in the large droplet.

**Figure 2.** (**A**) Schematic of a SRES-based immunoassay for PSA detection [52]. (**B**) Schematic of a SERS-based multiplex immunoassay detection for PSA and AFP [16]. Reproduced with permission from [52]. Copyright American Chemical Society, 2016. Reproduced with permission from [16]. Copyright Elsevier, 2018.

C-reactive protein (CRP) has been widely studied to reveal the relevance of inflammation and cardiovascular diseases. Guo et al. [56] developed an enzymatic activation strategy combined with immunoassay analysis to activate reduction caged Raman reporters for SERS detection of CRP. Agarose beads modified with capture antibodies, CRP and horseradish peroxidase (HRP)-modified detection antibodies were used to constitute the sandwich assay. The HRP activated reduction caged LMG to generate SERS active MG for measurements. A linear curve on a logarithmic scale was obtained between CRP concentrations and SERS signals. In addition to the immunoassay strategy, label-free strategy was also suitable for CRP detection [57]. 3D AgNPs aggregates functionalized by phosphocholine were constructed to selectively capture CRP and to enhance the SERS efficiency accordingly.

Alzheimer's disease was a chronic and progressive neurodegenerative disease causing dementia in the senior population. Amyloid-β peptide, a biomarker to diagnose Alzheimer's disease, was hydrophobic in nature. Distinct from the above disease biomarkers, label-free strategy was usually used in SERS detection of amyloid-β. Due to synergistic effects, many nanohybrids were designed as SERS-active substrates. For amyloid-β detection, hexagonal-packed lotus seedpod like array substrate [19], biomimetic lipid membrane [58] and SiNPLs@AgNPs [21] were constructed. 3D SERS platforms [59] were also designed to enhance the "hot spot" formation in 3D. A 3D GO based SERS substrate decorated by core-shell nanoparticle was developed from hybrid 2D GO cross-linked by amine-modified PEG. The SERS enhancement factor for 3D SERS substrate was around 3.9 × 10<sup>12</sup> with the LOD of 500 fg/mL amyloid-β.

Neuron-specific enolase (NSE) is a mortality predictor in traumatic brain injury patients. "Sandwich" immunoassay methods were generally used in SERS detection of NSE with different substrates. Wang et al. [60] utilized an indium tin oxide conductive glass slip with hollow gold nanospheres (HAuNPs) as the SERS substrates instead of normal silver nanomaterials. Nile blue A and NSE antibody functionalized HAuNPs were used as SERS nanotags to achieve the sensitive detection of NSE with the LOD of 0.1 ng/mL. Subsequently, lateral flow glass-hemostix (FGH) combined with Au nanocage was used as SERS substrates to detect NSE in blood plasma achieving a LOD of 0.74 ng/mL [61]. To realize rapid and low-cost clinical diagnosis, a paper-based SERS lateral flow strip was developed to detect NSE. Au nanostar@Raman Reporter@silica NPs was employed as SERS nanotag that exhibits superior performance compared to the colorimetric methods with the LOD of 0.86 ng/mL [62], indicating that paper-based SERS detection had a grea<sup>t</sup> potential to meet the requirement of point-of-care (POC) testing.
