**2. Results and Discussion**

#### *2.1. Detection of the Intact Ricin Molecule by MALDI-TOF MS*

Direct analysis of the samples through MALDI-TOF MS, led to a rapid identification of the characteristic peak of the intact molecule of ricin at *m*/*z* close to 64 kD, as shown in Figure 1. The exact position of this peak can change according to the simultaneous existence of different isoforms. The non-irradiated sample (0 kGy) showed a peak relatively intense in this region of the spectra (Figure 1), while the irradiated samples showed a reduction in intensity of this peak from 5000 a.u (non-irradiated sample) to around 1500 a.u (for sample irradiated at 10 kGy), and close to 0,000 a.u (for samples irradiated at 20 and 30 kGy). Table 1 reports the ratio signal/noise (S/N) for each sample, considering the triplicates analyzed. As can be seen, the irradiation dosage of 10, 20 and 30 kGy provoked average reductions of the signal to noise (S/N) of 65.9%, 91.8% and 97.5%, respectively, compared to the non-irradiated sample.

**Figure 1.** MALDI-TOF MS spectra of the non-irradiated ricin sample (0 kGy) and the samples irradiated at 10, 20 and 30 kGy.

**Table 1.** Ratio signal to noise (S/N) of ricin in the samples for each irradiation dosage (average values, three repetitions).


Direct analysis through MALDI-TOF MS showed useful for verification of the integrity of the ricin molecules present in the samples. The presence of a peak in the range between *m*/*z* 62,000 and 68,000 should be considered as an initial detection, indicating, in a semi-quantitative way, the possible presence of non-degraded ricin. This method has the advantage of being relatively rapid.

## *2.2. Unequivocal Identification of Ricin*

According to the procedures adopted by the Organization for Prohibition of Chemical Weapons (OPCW) [9], the unequivocal identification of a chemical substance related to the CWC should be made through two di fferent analytical methods, and one of them must be spectrometric. In order to meet this criterion, the identification of the ricin presence in the samples was performed through MALDI-TOF MS, using of peptide mass fingerprint (PMF) method after extraction of the protein separated through SDS-PAGE. Confirmation of the results was done with MALDI-TOF MS/MS.

## 2.2.1. SDS-PAGE in Reducing Conditions

Samples irradiated at 0, 10, 20 and 30 kGy were submitted to SDS-PAGE in reducing conditions. This technique allowed separation of the analytes of interest from other proteins and impurities. Figure 2 shows the picture of the polyacrylamide gel from the SDS-PAGE obtained after revelation with Coomassie blue. The hydrogenation promoted by the reducing agen<sup>t</sup> DTT caused the breaking of the disulfide bond between RTA and RTB. The main bands of interest were identified with numbers 1 and 2 in Figure 2 according to the crescent order of its respective molecular masses. Band 2, of higher molecular weight, presented approximately double of the length of band 1 in the vertical direction. This observation suggests the possibility of a superposition of the signals of two polypeptide chains, with similar molecular weights, at band 2.

**Figure 2.** SDS-PAGE gel of samples irradiated at 0, 10, 20 and 30 kGy. M = marker. The bands of interest were identified with numbers 1 and 2. They were removed from the gel and identified by MALDI-TOF/MS..

One single band from SDS-PAGE may contain more than one isoform which are di fferentiated from each other by the degree of glycosylation because di fferences in the contents of sugar generate small di fferences in mass [34]. Fultonet et al. [35] and Kim et al. [33], observed that RTA isolated and purified exhibit two bands in the gel revealed with Coomassie blue, while RTB does not present any heterogeneity. Therefore, the whole molecule of ricin splits into two protein bands. The upper band is a mixture of RTB and the first isoform of RTA, while the lower band corresponds exclusively to the second isoform of RTA [33]. All samples presented two bands in the SDS-PAGE gel. The sample irradiated at 30 kGy presented less intense color compared to the others. Considering that the intensity is related to the amount of protein bound to the colorant used, we can deduce that this sample contains the lower concentration of ricin.

In order to make a clear identification, the two bands were collected and analyzed by MALDI-TOF MS. Results (shown in the next topic) allowed the identification of band 1 as one of the isoforms of RTA and band 2 as a mixture between RTB and the second isoform of RTA.

#### 2.2.2. Ricin Identification through MALDI-TOF MS

The two bands in the region of interest in the SDS-PAGE gel were cut, discolored, and digested with trypsin. The peptides obtained this way were extracted from the gel and analyzed through MALDI-TOF MS. Figure 3 shows the result obtained for band 1 of the non-irradiated and the sample irradiated at 30 kGy for comparison purposes.

**Figure 3.** MALDI-TOF MS mass spectra of band 1 (RTA) of the ricin samples. (**A**) Non-irradiated and (**B**) irradiated with 30 kGy.

The mass standard spectrum, also known as the fingerprint of peptides, was used for the identification of the protein. At first, the list of the peak masses was exported to the program Biotools (Bruker®). After, through the search mechanism MASCOT PMF, the experimental results were compared to the information available in the data banks: SwissProtb [36] and National Center of Biotechnology Information (NCBI) [37]. The mass standard obtained from the analysis of band 1 showed to be similar to the chain RTA, with the probability of being a random event <0.05.

The radiolysis process may occur directly over the target molecule as a primary effect, or indirectly, through the formation and reaction of free radicals with other molecules present [38]. The gamma irradiation may denature proteins and reduce the amino acids content. Some residues, like the ones containing sulfur, can be more susceptible to radiolysis [39,40]. These phenomena may alter the intensity of peptides peaks as shown in Figure 3.

In Figure 3 the peaks were labeled in order of maintaining the correspondence with the peptides nomenclature shown in Table 2, which presents the expected masses for the complete proteolysis of ricin by trypsin. As not all theoretical scissions occurred, some peaks in Figure 3 were labeled as a sum of peptides, indicating that they remained connected. Table 2 lists the *m*/*z* values measured with the theoretical values of the corresponding peptides, and the respective amino acid sequences and the positions occupied in RTA.

Most of the peptides were identified without modifications, by the mass of their quasi-molecular ion [M + H]<sup>+</sup>, like, for example, A6, A9, A10, A11, A12, A14, A19 and A20.

The presence of some peaks suggests that the proteolysis was incomplete, producing one single peptide while two or three were expected. This happened for signals at *m*/*z* A1 + A2, A7 + A8, A10 + A11, A13 + A14, A16 + A17 and A16 + A17 + A18.

As the peptide A7 + A8 has an estimated mass of 4083.214 Da, its quasi-molecular ion [A7 + A8 + H]+ present *m*/*z* out of the range analyzed (700–3500) and, therefore, cannot be detected. However, it is expected that its doubled protonated ion at *m*/*z* = 2042.615 [A7 + A8 + 2H]<sup>2</sup>+ be detected. This is compatible with the signal observed at *m*/*z* = 2042.4 (see Table 2).


**Table 2.** Ricin peptides identified by MALDI-TOF MS in the band 1 of the non-irradiated sample.

a Value corresponding to the double protonated ion. b Considering the formation of the polyacrylamide adduct; c Considering oxidation of methionine to methionine sulfoxide.

The identification of peptide A13 was not trivial due to the presence of cysteine and methionine residues at positions 171 and 174, respectively. Cysteine residues can react with the acrylamide of the gel, adding 71.04 Da to the mass of the peptide. The signal observed at *m*/*z* = 1653.9 is compatible with this modification. The methionine residue may oxidize to methionine sulfoxide, originating one single peak related to A13. It was possible to identify the presence of this peak together with A14, as a low-intensity peak at 2806.6, corresponding to A13 + A14.

Peptide A14 was also associated with two other peaks at *m*/*z* 1172.2 and 1188.2 (less intense). The di fference of 16 units between them can be explained due to the oxidation of the methionine residue at position 188. The same phenomenon explains the peak of peptide A23 at *m*/*z* 2228.5.

Peptides A3, A4, A15, A16, A17, A18, A21 and A22 were not detected because their masses are <700 Da, and, therefore, out of the range analyzed (700–3500).

The most intense signals of the spectra (Figure 3) were identified as A10 and A9, in descending order. This result is important because these two peptides allow di fferentiating ricin from the lectin RCA120 also present in castor bean samples. The similarity between the amino acids of RTA and RCA120 is superior to 93%. By comparing the RTBs of both proteins, this value drops to the still high value of 84% [41]. So, in order to eliminate doubts in the identification, it is of fundamental importance to find intense signals of some peptides that allow di fferentiating ricin from RCA120, like A5, A7, A9, A10, A11, A13, A22, B14, B15, B18, B19 and B20 [42].

Table 3 shows a comparison of the peptide sequences of RTA and RCA120 between positions 86 and 124. The di fference between them is at amino acids 114 and 115, underlined in Table 3. While ricin holds an arginine (R) and a tyrosine (Y) at these positions, RCA120 holds a serine (S) and a phenylalanine (F). As trypsin works on the R, the cleavage happens only between amino acids 114 and 115 of ricin. Therefore, only ricin has the peptides A9 and A10, with masses 3307 and 1310 Da. The equivalent sequence of RCA120 has one single peptide with a mass of 4513 Da.

Due to the relevance of peptides A9 and A10, and the intensity of its peaks in the mass spectra of Figure 4, the ions at *m*/*z* 3307 and 1310, were chosen for confirmation of the sequence of amino acids by MALDI-TOF MS/MS as discussed in the next topic.


**Table 3.** Comparison between the amino acid sequences of RTA and RCA120 between positions 85 and 126 \*.

\* Different amino acids in both sequences are underlined. The digestion with trypsin leads to peptides A9 (in red) and A10 (in blue) in ricin but keeps RCA120 as a single peptide (in green).

MALDI-TOF MS analysis of band 1 of the SDS-PAGE gel identified the presence of RTA in the non-irradiated sample. As band 1 also shows up visible in the SDS-PAGE gels of the irradiated samples, it is reasonable to suppose that RTA is also present in these samples. In fact, this was confirmed by the mass spectra shown in Figure 3 where it is possible to identify the peaks of the selected ions A9 and A10 corresponding to RTA, contrasting with a single peptide for RCA120 (see Table 3).

The main peaks found in band 1 and identified as peptides of RTA, were also observed in the spectra obtained for band 2 of the SDS-PAGE gel for the non-irradiated sample, shown in Figure 4. The main di fference was the presence of peptides that compose RTB that were not observed before. Figure 4 and Table 4 present the results of the MALDI-TOF MS analysis of band 2 for the non-irradiated sample. The base ion, located at *m*/*z* 2230.9, corresponds to the peptide B13 (AEQQWALYADGSIRPQQNR).

**Figure 4.** MALDI-TOF MS mass spectra obtained for band 2 of the SDS-PAGE gel of the ricin samples. (**A**) Non-irradiated and (**B**) irradiated with 30 kGy. Magnification of the mass spectra in the region of peptides B1 and B6 is shown in the spectra of the non-irradiated sample.




**Table 4.** *Cont.*

a Considering the oxidation of methionine to methionine sulfoxide. \* Peptide of RCA120.

Besides B13, we also identified B1 and B6 among the peptides expected for RTB. These two, however, presented signals with low intensity when compared to the base ion and, therefore, are magnified in Figure 4. Other three peaks correspond to the clusters B6 + B7, B15 + B16 + B17 and B17 + B18 + B19. Peaks at *m*/*z* 1533.8 and 1889.1 do not belong to ricin but correspond to the peptides B6 \* and B18 \* of RCA120, being indicatives of a third component in the upper band of the SDS-PAGE gel (Figure 4 and Table 4).

The identification of the peptides present in band 2 was more difficult than for band 1. First due to the presence of different chains interfering in the spectra from each other, and increasing the complexity of the matrix. Besides, several peptides from RTB possess mass values below (B2, B4, B7, B8, B9, B15 e B17) or over (B12) the range of calibration for the method used, and, therefore, could not be identified separately. Finally, even after the addition of a reducing agen<sup>t</sup> (DTT) during sample preparation, some S−S bonds do not break and others may rebind naturally. Therefore, instead of producing one single peptide, many different combinations of fragments with different masses could have happened, making it difficult the identification. A usual alternative to inhibiting the formation of new disulfide bonds after sample reduction is the addition of an alkylating agent, like iodoacetamide or iodoacetic acid, which covalently binds to the thiol group of cysteine. In this case, one should consider the increase in mass due to the addition of this new group.

Despite its major complexity related to band 1, results make it clear that band 2 is composed by the superposition of signals from RTB and one of the isoforms of RTA, corroborating with the literature [33]. Additionally, we also found evidence of the presence of peptides from RCA120. The presence of this contaminant is justified because this is a natural protein of castor bean plants with chains similar to ricin.

Results of the MALDI-TOF technique analysis of the irradiated samples of band 2 (Figure 5), were very similar to the non-irradiated sample discussed before. The same peptides were identified, and the main difference observed was the intensity reduction of the signals in the spectra compared to the non-irradiated sample. These results show that the use of the technique of MALDI-TOF after separation through SDS-PAGE allowed identification of the presence of ricin in all samples studied, including the ones irradiated at 30 kGy.

## 2.2.3. Analysis by MALDI-TOF MS/MS

In order to confirm the identification of ricin by a second spectrometric technique, two peptides from RTA and one from RTB were verified by MALDI-TOF MS/MS. The first and second precursor ions selected were the ones with *m*/*z* 1310 Da and 3.307 Da, due to the high intensity of its peaks in the mass spectra, and the relevance of peptides A10 and A9 for the differentiation between ricin and RCA120. The third ion was the one corresponding to B13, with *m*/*z* 2231, for being the most intense related to RTB.

The MALDI-TOF MS/MS spectra corresponding to ion at *m*/*z* 1310 is shown in Figure 5. The data obtained from this spectrum were analyzed through the software Bruker Biotools® (Version 2.2, Bruker Daltonik GmbH, Bremen, Germany), together with the search mechanism MASCOT, and compared to the data banks SwissProt [36] and NCBI [37]. Results were compatible with the amino acids sequence YTFAFGGNYDR, confirming the identification of peptide A10 from ricin.

**Figure 5.** MALDI-TOF MS/MS spectra corresponding to the fragments of precursor ion *m*/*z* 1310 (**A**). Analysis of the MALDI-TOF MS/MS spectra of the precursor ion *m*/*z* 1310 (**B**).

Table 5 lists the fragments of peptide A10 identified by MALDI-TOF MS/MS. The following ions of series –y and –b of peptide A10, presented correspondence with the products generated from precursor *m*/*z* 1310: y1, y2, y3, y4, y5, y6, y7, y8 and y9; b2, b3 and b4. We also found some ions of the series a (a1, a2 and a7) and immonium, which contributed to reinforcing the interpretation of the results.


**Table 5.** Fragments of peptide A10 identified by MALDI-TOF MS/MS.

Figure 6 presents the MALDI-TOF MS/MS spectra corresponding to the fragments of ion *m*/*z* 3307, and the corresponding analysis through the software Bruker Biotools® together with the search mechanism MASCOT, and compared to the data banks SwissProt [36] and NCBI [37]. Like before, it is possible to verify that the results are compatible with the amino acid sequence AGNSAYFFHPDNQEDAEAITHLFTDVQNR, confirming the identification of the peptide A9 of ricin. As shown in Table 6, the following series y and b were found: y1, y3, y4, y5, y6, y7, y8, y9, y11, y12, y13, y14, y15, y16, y17, y18, y20, y21, y22, y25 and y26, together with b3, b4, b7, b9, b12, b15, b23 and b25.

Finally, in order to definitely identify ricin in the samples, the last ion selected for the analysis by MALDI-TOF MS/MS was the *m*/*z* 2231. We tried to verify if the products formed would be compatible with the fragments of peptide B13. Results are shown in Figure 7 and Table 7.

**Figure 6.** MALDI-TOF MS/MS spectra corresponding to the precursor ion *m*/*z* 3307 (**A**). Analysis of the MALDI-TOF MS/MS spectrum of the precursor ion *m*/*z* 3307 (**B**).



The presence of ions y1, y5, y9 and y14, together with B3, B4, B6, B14 and B18, allowed confirming the similarity between the peak observed in the spectra of *m*/*z* 2331 and the sequence of amino acids of peptide B13 (AEQQWALYADGSIRPQQNR).

**Figure 7.** MALDI-TOF MS/MS spectrum corresponding to the fragments of the precursor ion *m*/*z* 2231. (**A**) Analysis of the MALDI-TOF MS/MS spectrum of the precursor ion *m*/*z* 2231 (**B**).


**Table 7.** Fragments of peptide B13 (AEQQWALYADGSIRPQQNR) identified by MALDI-TOF MS/MS.

#### 2.2.4. Determination of the Toxic Activity by MALDI-TOF MS

The active site responsible for the toxicity of ricin is located in RTA between residues Tyr80 and Trp211. The residues playing the most important role in the mechanism of adenine removal from rRNA 28S are Tyr80, Tyr123, Glu177 and Arg180 [20,43]. This information together with the PMF spectrum obtained by MALDI-TOF MS (Figure 3) allows correlating the ricin activity to peptides A8 to A13. Among them A8, A10 and A13 are the ones containing the most relevant residues [20,43].

All peptides in the region of the active site of ricin were identified by MALDI-TOF MS for both the non-irradiated and the irradiated samples (Figure 3). The presence of these peptides suggests the possibility of toxic activity even in the samples irradiated at 30 kGy.

In order to confirm whether the samples presented toxic activity, a non-irradiated sample and another irradiated at 30 kGy, were incubated with a buffer solution containing DNA substrate with a nucleotide sequence similar to rRNA 28S. A buffer solution containing only the DNA substrate was used as a control. Aliquots were collected at three different times of incubation (0, 4 and 24 h) and analyzed by MALDI-TOF MS. Results are shown in Figure 8. At the beginning of the reaction (letters "a", "b" and "c" in Figure 8), all samples presented a unique set of intense peaks with *m*/*z* values starting at 3697, followed by 3719. These spectra are compatible with the mass of the intact oligonucleotide (GCGCGAGAGCGC) (Figure 8). The first signal corresponds to the quasi-molecular ion [M + H]+ and the others to adducts of salts usually present, like sodium salts [M + Na]<sup>+</sup>.

After 4 h of reaction, no alteration was observed in the control sample. However, in the samples incubated with ricin (0 and 30 kGy), it was observed a peak at *m*/*z* 3564 with very low intensity compared to the base peak [M + H]+ (Figure 8d,e,f). This same peak became much more intense in the aliquots collected after 24 h of incubation with ricin, reaching around 70% of the intensity of the

base peak [M + H]+ (Figure 8h,i). The control sample presented only the set [M + H]+ and its adducts (Figure 8g). The difference of *m*/*z* between the quasi-molecular ion [M + H]+ and the peak at 3564 is of 133 units. This is compatible with the replacement of one adenine base of the nucleotide sequence by a hydrogen atom. The label [M − A + H]+ was used to identify this peak in Figure 8.

**Figure 8.** Verification of the toxic activity of ricin by MALDI-TOF MS. (**a**) Control at 0 h; (**b**) Sample 0 kGy at 0 h (**c**) Sample 30 kGy at 0 h; (**d**) Control after 4 h; (**e**) Sample 0 kGy after4h(**f**) Sample 30 kGy after 4 h; (**g**) Control after 24 h; (**h**) Sample 0 kGy after 24 h (**i**) Sample 30 kGy after 24 h.

It was possible to see by MALDI-TOF spectrometry that both samples, the non-irradiated and irradiated at 30 kGy, attacked the DNA substrate, provoking the removal of the adenine nucleotide from the sequence GCGCGAGAGCGC. This result is compatible with the MALDI-TOF MS spectra of the samples where we had already identified the presence of peptides related to the active site of ricin (Figures 3 and 4) and shows that irradiation at 30 kGy is not enough to eliminate totally the toxic activity of ricin.
