Self-assembled GA-Repeated Peptides as a Biomolecular Scaffold for Biosensing with MoS2 Electrochemical Transistors

Biosensors with two-dimensional materials have gained wide interest due to their high sensitivity. Among them, single-layer MoS2 has become a new class of biosensing platform owing to its semiconducting property. Immobilization of bioprobes directly onto the MoS2 surface with chemical bonding or random physisorption has been widely studied. However, these approaches potentially cause a reduction of conductivity and sensitivity of the biosensor. In this work, we designed peptides that spontaneously align into monomolecular-thick nanostructures on electrochemical MoS2 transistors in a non-covalent fashion and act as a biomolecular scaffold for efficient biosensing. These peptides consist of repeated domains of glycine and alanine in the sequence and form self-assembled structures with sixfold symmetry templated by the lattice of MoS2. We investigated electronic interactions of self-assembled peptides with MoS2 by designing their amino acid sequence with charged amino acids at both ends. Charged amino acids in the sequence showed a correlation with the electrical properties of single-layer MoS2, where negatively charged peptides caused a shift of threshold voltage in MoS2 transistors and neutral and positively charged peptides had no significant effect on the threshold voltage. The transconductance of transistors had no decrease due to the self-assembled peptides, indicating that aligned peptides can act as a biomolecular scaffold without degrading the intrinsic electronic properties for biosensing. We also investigated the impact of peptides on the photoluminescence (PL) of single-layer MoS2 and found that the PL intensity changed sensitively depending on the amino acid sequence of peptides. Finally, we demonstrated a femtomolar-level sensitivity of biosensing using biotinylated peptides to detect streptavidin.


3: Device Fabrication of MoS 2 FET
The MoS2 field effect transistors (FETs) were fabricated as following steps. First, electrodes were fabricated with a standard photolithography technique by depositing 10-nm Ti and 40-nm Au in a vacuum thermal evaporator. MoS2 was grown on a Si substrate by chemical vapor deposition.
Then, the MoS2 were transferred on a Si wafer (with 270-nm thick SiO2 layer) with patterned electrodes. In this process, polystyrene (PS) was used as a supporting film and it was transferred 5 by a Poly(dimethylsiloxane) PDMS stamp 2 . The transferred PS supporting film was removed by immersing in toluene at 90 °C for 30 minutes. Sequentially, the substrate was rinsed with toluene at room temperature to remove the polymer residues completely at room temperature, and it was dried with nitrogen blow. As the result of this process, the CVD-grown MoS2 was placed on desired electrodes to form a FET. As a protection layer for electrodes, polymethyl methacrylate (PMMA) was deposited on electrodes with a toothpick by hands. The procedure is summarized in Figs. S3a and S3b. Raman spectrum of the MoS2 reveals two peaks at 383.7 cm -1 and 404.7 cm -1 as shown in Fig.S4c.
It is well known that these peaks correspond to the vibrational modes of E2g and A1g, respectively, and MoS2 has a single layer when the position difference between the peaks is 20 cm -1 or less 4,5 .
In this work, the difference of the peak position was 18.7 cm -1 . These observations support that MoS2 in FETs in this work consists of single layer.  We found that the current shows a small hysteresis. We also evaluated the threshold for both forward and backward sweeps in the same manner. We defined the hysteresis of the threshold voltage as the difference of these threshold voltages. Furthermore, we also derived a transconductance of FETs, which was obtained from a linear fitting as well. In Figs

5: PL measurements
The PL of single-layer MoS2 was obtained by an inverted microscope (Olympus IX73) with a spectrometer (Oxford instruments, Shamrock 193i) equipped with an electron multiplying charge coupled device (Oxford instruments, iXon-Ultra 888 EMCCD). The excitation light from a mercury lamp was guided to a sample through a band pass filter, a dichroic mirror, and 100X objective lens (N.A = 0.95). The 546-nm line in the mercury lamp was used for the excitation.
The spectrometer contains gratings and a mirror. These gratings and mirror are switchable. The mirror allows us to obtain PL images by fully opening an optical slit in the spectrometer (Fig.   S6a). Alternatively, the grating and optical slit allows us to measure spectrograph containing spatial information (Fig. S6b). This image shows local PL from the CVD-MoS2, which allows us 8 to see the spatial distribution of PL spectra over the single-layer MoS2. A typical PL spectrum is shown in Fig. S6c. The PL peak position and intensity varied at each location. It is probably due to the internal strain and cracks in the MoS2. The peak positions and intensity at each location were collected, and then the difference between before and after peptide assembly was calculated (Fig. 3).

7: Co-assembly of peptides
We defined the mixing ratio as a fraction of Bio-Y5Y concentration over total concentration of all peptides (Bio-Y5Y + QY5). All the sample showed a high coverage (Fig. S7). It indicates that the mixing of Bio-Y5Y did not affect their binding affinity to the MoS2 surface. In AFM images, some bare areas with round shape were found. It could be caused by air bubble at the interface when we incubated the peptide solution on the MoS2 surface. The thickness of monomolecular structures is uniform, indicating that Bio-Y5Y and QY5 are miscible each other when they selfassembled on surface.
9 Figure S7. AFM images showing the morphology of co-assembly of Bio-Y5Y/ QY5 with various mixing ratios.
We monitored the morphology of co-assembled peptides at the same location by in-situ AFM while we sequentially added streptavidin (SA) with different concentrations (Fig. S8). Before adding SA, we observed fine linear structures of peptides. It manifests that the co-assembled peptide forms long-range ordered structures even after mixing two kinds of peptides, Bio-Y5Y and QY5. It indicates that these peptides are miscible, probably because they shared the GA domain for their interpeptide interactions. After adding SA, we found that there are small dots with slightly bright color in height images increased as we increased the concentration of SA.
This tendency was similarly observed in phase images. It is perhaps because of binding of SA to biotin in the co-assembled peptides. The tendency of SA binding is consistent with our finding in the electrical measurements (Fig. 4).

8: Real time detection of streptavidin by MoS 2 biosensor
The source-drain current was measured by fixing the gate voltage and the source drain voltage constant at 0.5 V and 30 mV, respectively. After incubation of peptide with 50% mixing ratio, we waited for the source-drain current stabilized. Before adding SA in the solution, we replaced the solution with a 10-mM PB solution using a pipette several times to remove excess peptides.
After waiting for about 20 min, SA solutions with 10-mM PB were added every 10 minutes to characterize the response of MoS2 FET against the SA. The concentrations of SA were increased from 1 fM to 100 nM.