Molecular origin of the Raman signal from Aspergillus nidulans conidia and observation of fluorescence vibrational structure at room temperature

Successful approaches to identification and/or biological characterization of fungal specimens through Raman spectroscopy may require the determination of the molecular origin of the Raman response as well as its separation from the background fluorescence. The presence of fluorescence can interfere with Raman detection and is virtually impossible to avoid. Fluorescence leads to a multiplicity of problems: one is noise, while another is “fake” spectral structure that can easily be confused for spontaneous Raman peaks. One solution for these problems is Shifted Excitation Raman Difference Spectroscopy (SERDS), in which a tunable light source generates two spectra with different excitation frequencies in order to eliminate fluorescence from the measured signal. We combine a SERDS technique with genetic breeding of mutant populations and demonstrate that the Raman signal from Aspergillus nidulans conidia originates in pigment molecules within the cell wall. In addition, we observe unambiguous vibrational fine-structure in the fluorescence response at room temperature. We hypothesize that the vibrational fine-structure in the fluorescence results from the formation of flexible, long-lived molecular cages in the bio-polymer matrix of the cell wall that partially shield target molecules from the immediate environment and also constrain their degrees of freedom.

The total signal S recorded by a spectrometer can be expressed with respect to the excitation frequency as ( , ) = ( , ) + ( , ) + ( ) , where R, L, and B represent the Raman signal, fluorescence signal, and background signal, respectively.
Equation (4) then takes the form: where the prime denotes the derivative with respect to frequency . In Eq. (7), higher-order terms of ∆ are neglected, since ∆ itself is assumed to be small. The Raman signal can therefore be reconstructed as We note that the above method is equivalent to "Difference-Integration-δ_Deconvolution" [1].

Appendix 2: RodA hydrophobin
The gene rodA of A. nidulans encodes a small, moderately hydrophobic polypeptide [3]. Classified as a hydrophobin protein, the polypeptide is involved in the formation of filamentous rodlets on the surface of the A. nidulans conidia [3][4][5][6]. The rodlet nanostructure is significantly hydrophobic and allows for more efficient dispersal of the conidium throughout the local environment [3,4]. The deletion of the rodA gene abolishes the production of the corresponding polypeptide, which in turn inhibits the formation of the hydrophobic protein nanostructure on the surface of each conidium, and results in a comparatively smooth surface that is much less hydrophobic [3][4][5][6]. A. nidulans strains containing this deletion are termed ∆rodA, while their unmodified counterparts (which are otherwise considered isogenic with the mutants) are termed rodA + .
There are two commonly used configurations in ECDL applications: Littrow configuration and Littman-Metcalf configuration [7,8]. Taking into account both output efficiency and the requirement of a fixed output beam propagation-direction for varying wavelength, we abandoned the former option and chose the latter geometry. Our homemade ECDL consisted of a laser diode (L785P090, Thorlabs) operating at ~785 nm, a collimating lens, a diffraction grating, and an end mirror (see left panel in Figure S1). The zero-order diffracted beam served as the output, while the first-order diffracted beam was reflected back into the diode as feedback. The output wavelength could be continuously tuned by adjusting the angle of the end mirror on the first-order diffracted beam. The linewidth and central wavelength were also diagnosed. Laser linewidth was monitored by an oscilloscope, which was connected to a scanning confocal cavity module including a scanning Fabry-Perot interferometer (SCC-2500, VitaWave) and a scanning cavity driver (CFPD200, VitaWare) ( Figure S1, middle panel). The central wavelength was measured by a wavemeter (WA-1500, Burleigh Instruments Inc.) ( Figure S1, right panel).
We selected four wavelengths with a spacing of ~5 cm -1 . The frequency stability and spectral linewidth measurements were conducted simultaneously once a desired wavelength was achieved.
In order to ensure the stability of the homemade laser during spore measurements, the characterizations for each wavelength lasted at least one minute. Table S1 summarizes the results of the frequency stability and spectral linewidth analyses of our homemade ECDL. The laser had a linewidth less than 150 MHz, and its frequency stability was within 150 MHz. It should be noted that Raman spectroscopy requires a light source with frequency stability better than 1 cm -1 (or 30 GHz) as well as a linewidth narrower than 30 GHz [7]. Therefore, our homemade ECDL was sufficient for Raman applications. We also investigated the long-term stability of the laser (see the last two rows in Table S1). The laser was inspected after 7 and 20 hours, respectively, of continuous operation. Although the central wavelength drifted slightly (which otherwise might result in serious consequences, such as spurious Raman shift and line broadening [7]), this small change (0.03 cm -1 ) could be ignored in short one-minute detection acquisitions. These results confirmed that the tunable laser was stable and thus suitable for shifted excitation Raman difference spectroscopy (SERDS).

Appendix 4: Substrate
In order to determine the physical origin of the minor peaks in the measured raw spectra that were insensitive to slight changes in the excitation frequency, we compared magnified spectrum amplitude of substrate and spectrum amplitude of a single green rodA + spore (shown in Figure  S2(a)). A similar baseline subtraction using the AsLS method was also used to determine if there was any fine-scale structure beyond 1700 cm -1 (shown in Figure S2(b)). The substrate did not exhibit the same behavior evident in spores, even though its signal was magnified by two orders of magnitude, implying that the fine-scale features in the raw spectra of sample conidia do not originate from any systematic instrumental response but rather from the light-matter interaction between the laser excitation and the molecules within the conidium itself.  Figure S3(a) and (b) show the measured raw data and corresponding normalized spectra of a green rodA + spore, respectively. Figure S3(c) is the SERDS spectrum with fluorescence removal. It was obtained by subtracting one spectrum from the other in Figure S3(b). The corresponding reconstructed spectrum is displayed in Figure S3(d). The relatively featureless region beyond 1700 cm -1 implies that there are no Raman bands in this range.  Figure S4(a) and (b) displays the average raw data and normalized raw data of 100 yellow rodA+ spores, respectively. Figure S4(c) shows the SERDS spectrum of yellow rodA + spores free from fluorescence signals. Figure S4(d) is the retrieved Raman spectrum obtained by integrating the curve in Figure S4(a). The relatively featureless region beyond 1700 cm -1 implies that there are no Raman bands in this range, which is also consistent with the SERDS spectrum.