2.1.

Femtosecond Laser Set-Up

A 40-mJ Ti-sapphire laser (Coherent Inc., Santa Clara, California) at 800-nm wavelength was used to irradiate the liquid samples at two different pulse widths [full-width half-maximum (FWHM)]: $\sim 40$ or 10 ns. The laser gain medium is Ti-doped sapphire. The femtosecond laser source is a Kerr lens mode-locked oscillator (Mantis, Coherent Inc., Santa Clara, California)¹⁹ that seeds a chirped-pulse amplification system²⁰ composed of a grating stretcher, regenerative amplifier, multi-pass power amplifier, and grating compressor (Hydra, Coherent Inc., Santa Clara, California). The FWHM duration of the measured single-shot autocorrelation signal is 56 fs, corresponding to the time duration of 39 fs for an ideal Gaussian temporal profile (Fig. 1). From the extended “wings” of the autocorrelation signal, it is evident that the typical pulse does not have a truly Gaussian shape (due to, e.g., spectral broadening and high-order dispersion in optics such as the $f=1\text{m}$ lens). Blocking the oscillator and allowing the regenerative amplifier to Q-switch resulted in 10-ns pulses at the same wavelength. Variable doses—from a single laser pulse up to a 1000 pulses—were delivered to a cuvette at 1- to 10-Hz rep rate. The estimated laser fluence for both the femtosecond (short) and nanosecond (long) pulse was $0.5\mathrm{J}/{\mathrm{cm}}^2$, as determined by patterns (3.5-mm-beam diameter) on laser burn paper and a pyroelectric energy meter (Coherent LabMax, Santa Clara, California). The laser pulse was weakly focused with a 1-m focal length lens into the sample irradiation chamber. At the entrance to the chamber, the femtosecond and nanosecond laser irradiance are $10\mathrm{TW}/{\mathrm{cm}}^2$ and $40\mathrm{MW}/{\mathrm{cm}}^2$, respectively. Approximately 1% or 35% of the overall nanosecond or femtosecond laser pulse energy, respectively, was absorbed by water for the specified sample (10-mm optical length), as determined by energy meter measurements. This indicates a nonlinear absorption mechanism for the femtosecond laser pulse (vide infra).

Fig. 1

Single-shot femtosecond laser pulse shape.

2.2.

Sample Preparation and Irradiation

All chemicals were commercially available, and were obtained from Sigma Aldrich, St. Louis, Missouri. All solvents were purchased from Thermo Fisher Scientific, Waltham, Massachusetts, at LC-MS grade or higher. Bovine ubiquitin samples were dissolved in water at 20-μM concentrations. Bacillus globigii (B. globigii, also labeled Bacillus atrophaeus) spores were obtained from the US Army Laboratories (Dugway Proving Ground, Utah) and dissolved in plasma grade water at a concentration of $1\mathrm{mg}/\mathrm{mL}$. The matrix-assisted laser desorption/ionization (MALDI) matrix, $\alpha$-cyano-4-hydroxycinnamic acid, CHCA, (Fluka Analytical), was prepared at $10\mathrm{mg}/\mathrm{mL}$ in acetonitrile/0.1% trifluoroacetic acid (TFA) (70/30, v/v).

Typically, 55 μL of dissolved protein or suspended spore samples were transferred into a narrow, optically transparent quartz cuvette with internal dimensions of $2\mathrm{mm}(\mathrm{W})\times 10\mathrm{mm}(\mathrm{L})\times 4\mathrm{cm}(\mathrm{H})$ (International Crystal Laboratories, Garfield, New Jersey). The filled cuvettes were placed in the collimated beam (typical cross–section beam diameter 2 mm) for irradiation. These parameters ensure that the entire sample volume is irradiated with a single laser shot. The spore suspension was manually shaken after every 100 shots to avoid spore sedimentation. Irradiated aliquots were taken at different exposure times for subsequent MS analysis. A schematic of the irradiation set-up is given in Fig. 2.

Fig. 2

Experimental laser irradiation set-up.

2.3.

Sample Analysis

Irradiated and nonirradiated (control) ubiquitin samples were diluted in $50:50$ (methanol:water with 2% acetic acid) and loaded in nanospray emitters (New Objective, Woburn, Massachusetts). The samples were electrosprayed in positive ion mode and analyzed by high resolution MS and MS/MS on an LTQ-Orbitrap Velos (Thermo Fisher Scientific, Waltham, Massachusetts). Tandem MS and standard bioinformatics software were employed to characterize the modification site of the oxidized protein.

Positive ion MALDI mass spectra were acquired on an Autoflex tandem time-of-flight (TOF/TOF) instrument (Bruker Daltonics, Billerica, Massachusetts). Ions were desorbed by a ${\mathrm{N}}_2$ laser (337 nm). Standard accelerating voltage (20 kV) and delayed extraction (150 ns) conditions were used in positive ion linear mode.²¹ The mass spectrometer was calibrated using peptide and protein standards (average mass error in linear TOF mode $<3\mathrm{Da}$ across the mass/charge range below 10,000). For MALDI MS analysis, 0.5 μL of the protein solution was pipetted onto the MALDI stainless steel target, followed by an equal volume of CHCA. Prior to matrix addition, B. globigii spores were treated on the target with 0.5 μL of 10% TFA.²² The laser beam was manually rastered across the entire sample (3-mm diameter). For each spectrum, several hundred individual traces from single laser shots were accumulated, averaged and processed using the software provided with the instrument.

For scanning electron microscopy (SEM), both laser-irradiated and control spore suspensions were deposited on microscope grids and allowed to dry before analyzing them on an Hitachi S4700 scanning electron microscope (Hitachi High Technologies America, Dallas, Texas).

3.1.

Intact Protein Irradiation

It is well established that femtosecond laser excitation of dielectric materials, such as proteins and other biologically related species, differs markedly from nanosecond laser excitation.²³ Recently, we reported on a novel effect in the interactions of femtosecond lasers with biomolecules in water—oxidation of solvent-accessible sites (e.g., amino acids in proteins).²⁴ This is an entirely femtosecond laser-induced oxidation/reduction chemistry effect that is not observed upon nanosecond laser irradiation of the same sample at the same wavelength and energy density (Fig. 3). For example, upon femtosecond laser irradiation of ubiquitin (a small 76 amino acid protein with an average molecular weight of 8565 Da) regular peaks uniformly spaced 16 Da above the intact molecular ion peak are observed in its mass spectrum. Up to 14 oxidized sites per ubiquitin molecule are clearly discerned from these characteristic mass shifts based on high resolution MS data (Fig. 3). Oxidation occurs even upon irradiation with a single femtosecond laser pulse. Under our experimental conditions, intact ubiquitin is $>60\%$ oxidized after 100 laser pulses. The femtosecond laser–water interaction is akin to the interactions of ionizing radiation with water. As a result of gamma-iradiation²⁵ or electron-irradiation²⁶ of solutions, hydroxyl radical species are formed and subsequently interact with solvent-exposed groups of biomolecules. Similarly, femtosecond laser interactions with water induce the formation of hydroxyl radicals and solvated electrons. Although the photon energy of the femtosecond laser is 1.55 eV, compared to the 6.5-eV ionization energy of a water molecule in liquid,²⁷ the incident laser intensity of $10\mathrm{TW}/{\mathrm{cm}}^2$ is sufficient to enable 5-photon ionization of the water. For this laser intensity, multi-photon ionization has been found to be the dominant free-electron generation mechanism, over tunneling ionization and strong-field ionization (transition to tunneling ionization occurs at slightly higher irradiances of 13 to $26\mathrm{TW}/{\mathrm{cm}}^2$).²⁸ Subsequent fast (sub-microsecond time scale) secondary reactions with diffusion-limited rates result in accessible side chains of the proteins that are oxidized or reduced. Tandem MS has been utilized to discern the modified amino acid(s) for singly oxidized ubiquitin.²⁴ For a singly oxidized ubiquitin molecules, it is clear that oxidation occurs only on the $N$-terminal amino acid Methionine (Met). This result is in agreement with the reported much higher Met reactivity towards hydroxyl free radicals compared to other amino acids¹⁸ as well as the free and water-exposed $N$-terminal of the molecules. However, for doubly and more highly oxidized ubiquitin, we cannot rule out oxidation sites that are variable across the chain. A very promising application of this femtosecond laser technique for hydroxyl radical generation is for structural biology “foot-printing” of exposed amino acid side chains.¹⁸ Till now, foot-printing has typically required complex and orders of magnitude more expensive equipment than lasers (e.g., synchrotrons¹⁸), or the presence in the solution of non-native reagents (e.g., hydrogen peroxide²⁹) that could perturb the native biomolecular structure.

Fig. 3

Comparison of electrospray ionization mass spectra of ubiquitin after irradiation with: (a) 100 shots with the nanosecond laser, and (b) 100 shots with the femtosecond laser. Oxidized ubiquitin with up to 14 oxidation sites is observed after femtosecond laser irradiation.

3.2.

Intact Bacillus Spore Irradiation

Bacterial spores, formed when microorganisms encounter nutrient shortages or are exposed to other adverse factors, are some of the most refractory life forms. B. globigii (B. atrophaeus) spores have been used for decades as a simulant for the biowarfare agent Bacillus anthracis.³⁰ Electron micrographs of B. globigii spores, obtained before and after exposure of the spore suspension to femtosecond laser irradiation, are shown in Fig. 4. Clearly, the spores have completely disintegrated after irradiation at the current experimental conditions. This effect is in stark contrast to other types of energy deposition to intact spores, e.g., prolonged spore exposure to high voltage high frequency atmospheric plasma.³¹ During ultrafast excitation of condensed matter with the femtosecond laser pulse, the irradiated volume cannot transfer significant amounts of the absorbed energy to the surrounding material since phonon relaxation times in condensed matter are $\sim 100\mathrm{fs}$. Consequently, during the excitation process, electrons absorb the laser pulse energy and release it to the surrounding material at rates characteristic of the material (related to the electron-phonon coupling constant). During these early stages of excitation, various material effects occur. A “direct hit” of the micrometer scale length spores and subsequent phase explosion⁵ may lead to their complete disintegration. We speculate that microscale mechanical shock waves² in water may also cause complete spore disintegration and ultimately their inactivation. Observation of macroscopic bubbling in bacterial suspensions after 10- to 100-fs laser shots lends credence to this notion. Similar bubbling is not observed under our experimental conditions in the fully aqueous protein samples. Macroscopic bubble formation is likely due to spore suspension components that have stronger optical absorption bands and lower ionization thresholds compared to that of pure water.²⁸ For example, a decreased ionization threshold leads to more rapidly ionized plasma that can seed further collisional ionization.² The dense plasma electrons transfer energy to the heavier ions, resulting in thermoelastic stress-induced shock waves and bubble formation.²⁸

Fig. 4

Bacillus globigii spore micrographs (magnification $50,000\times$, typical spore length is 1 μm): (a) control and (b) after 500 laser shots with the femtosecond laser.

We have examined by MALDI TOF MS the biomarker signatures, characteristic of B. globigii spores before and after irradiation—Fig. 5. The predominant species, detected in the MALDI MS of intact spores are the SASP that comprise about 20% of the dry spore weight.²¹ For B. globigii spores, the three major SASP have average molecular weights of 7068 Da (SASP1); 7333 Da (SASP2), 8890 Da ($\gamma$-SASP).^21,22 As a result of femtosecond laser irradiation, the SASP2 is quickly oxidized, evidenced by the appearance of satellite peaks 16 Da above the unmodified peak. The degree of oxidation correlates with the number of laser pulses. This effect is similar to the SASP oxidation, observed upon rapid on-target oxidation after spore treatment with 30% hydrogen peroxide solution.²² In this latter experiment, predominantly the Met residues found in the SASPs are oxidized. The masses of Met-containing SASPs increase by $N\times 16\mathrm{Da}$, where $N$ is the number of Met residues per protein. In parallel to this on-target ${\mathrm{H}}_2{\mathrm{O}}_2$ oxidation, the SASP2 containing two Met is mostly oxidized after femtosecond laser irradiation as well. In contrast, the $\gamma$-SASP, which does not contain any Met residues, remains unchanged. These observations suggest that a chemical oxidation reaction takes place as a result of hydroxyl radical formed by the femtosecond laser pulse, similar to what is observed during irradiation of individual proteins (vide supra). Presumably, the mechanical spore destruction releases intact SASP, which are subsequently oxidized upon femtosecond laser irradiation. No oxidation or spore disruption is observed even after irradiation with 1000 pulses by a 10-ns laser at the same wavelength and fluence.

Fig. 5

Comparison of positive ion matrix-assisted laser desorption/ionization mass spectra of B. globigii in the small acid-soluble spore protein mass range; (a) control, (b) after 10 shots with the femtosecond laser, (c) after 500 shots with the femtosecond laser, and (d) after 500 shots with the nanosecond laser.

In a recent study on the interactions of ultra-short pulses (90 fs at 800 nm) with biological molecules and cells, no effects have been observed.¹⁶ In those experiments, the irradiation times are on the order of 1 h, while the individual laser pulse energy is only on the order of 1 nJ ($100\mathrm{MW}/{\mathrm{cm}}^2$ on target power density). The integrated incident laser energy for that experiment is $\sim 288\mathrm{J}/\mathrm{mL}$ (2-mL total irradiated volume), which is almost 600 times higher than the total dose in our experiment ($0.5\text{-}\mathrm{J}/\mathrm{mL}$, 50-μL sample volume). We point out that the laser power density for our experiment is five orders of magnitude higher, which indicates that the power density (peak power) rather than total incident energy determines the types of effects for femtosecond laser interactions with biological matter. Thus, effects of strand breaks and uncoiling of DNA plasmids in aqueous media are observed upon irradiation with an 820 nm, 45-fs pulse laser at power densities up to $12\mathrm{TW}/{\mathrm{cm}}^2$.³² Similarly, almost instant skin lesions have been observed in Yucatan mini-pigs irradiated with 810 nm, 44-fs terawatt laser pulses (8.2-mJ total deposited energy for 12-mm spot size).³³