In vivo citrullination is known to have low stoichiometry [14, 115]. As discussed earlier, proteomics analysis of complex samples including synovial fluid often results in a small number of citrullinated peptides given the number of analysis hours. With the increased interest in large-scale identification of citrullination sites and the low number of citrullinated peptides routinely identified in proteomics experiments, efficient enrichment is essential.
In 2010, Tutturen et al.  developed an enrichment technique for citrullinated peptides. Briefly, using a multistep process, sarcosine dimethylacryla- mide resin (PL-DMA) is functionalized with ethylenediamine (Figure 7.6). To this, 4-hydroxymethylbenzoic acid is added, followed by bromoacetic acid. Finally, 4-hydroxyphenylglyoxal is added to produce citrulline reactive beads (CRBs). The citrulline residue covalently binds to the CRB, and noncitrulli- nated peptides are washed from the beads. The bound peptides are released from the beads by washing with NaOH and guanidine. Peptides that were bound to the CRBs have a mass increase of 190.03 Da caused by the glyoxal derivative on the CRB. The technique was initially tested by enriching two synthetic citrullinated peptides (AcRSSVPGVR and SAVQAcRSSVPGVR). The resulting enriched sample was analyzed by MALDI-TOF MS, with both peptides successfully enriched compared to unmodified peptides.
Figure 7.6 A schematic of citrulline reactive bead (CRB) enrichment. (a) The structure of the CRB. (b) Binding citrullinated peptides to the CRB, washing unmodified peptides from the beads, and elution of citrullinated peptides bound to the glyoxal derivative. (c) The structure of the citrulline plus glyoxal derivative. Source: Tutturen, 2010. . Reproduced with permission from Elsevier.
Figure 7.6 (Continued)
To assess the technique further, MBP (human) was treated with PAD4, digested with Lys-C, and enriched. The eluted peptides were analyzed by MALDI-TOF MS. After enrichment, six peaks were observed and identified as five citrullinated peptides (one peptide was identified as both singly and doubly citrullinated; see Figure 7.7). One of the drawbacks of this method is the background signal observed from substances derived from the polymer of the beads (Figure 7.7a).
An alternative enrichment method was developed in 2013  by the same group. Initially, 4-glyoxalbenzoic acid is synthesized by oxidizing 4-acetyl- benzoic acid, which is then reacted with amine-PEG2-biotin to produce biotin-PEG-GBA (BPG). This tag can be enriched using streptavidin beads (Figure 7.8). The authors suggest mixing the biotin tag with the peptide mixture and then using SCX chromatography to separate any unreacted tags from the peptide mixture. The peptide mixture is mixed with the streptavidin beads, and the noncitrullinated peptides are removed. The tagged peptides are eluted from the streptavidin beads using excess biotin. The resulting tagged peptides have a mass addition of 516.4 Da. CID of peptides containing a BPG tag produces an intense fragment ion from the BPG at m/z 270.13, a potential product ion to either trigger additional fragmentation  or for use as a marker for modified peptides, giving greater confidence to the citrulline assignment. As in the previous method, MBP was citrullinated with PAD4, digested with Lys-C, and enriched. Analysis was performed by MALDI-TOF. The five citrullinated peptides identified in the previous work were identified; however, the peptide RPSQRHGSK (identified as two singly citrullinated peptides and one doubly citrullinated peptide with CRB enrichment) was only identified with R5 being citrullinated. One additional citrullinated peptide was identified (cRGSGK). This enrichment method appears cleaner than the CRB method as there is no polymer contamination; however the production of the BPG tag requires multiple steps.
Tutturen et al.  have applied the BPG enrichment method to a complex synovial fluid sample from a patient with RA to assess the citrullinome. They compared LC-MS/MS analysis of samples with and without enrichment. The supernatant sample was split three ways: one analyzed without purification,
Figure 7.7 The enrichment of citrullinated peptides using the CRB method. (a) Citrulline enrichment of myelin basic protein using citrulline reactive beads: unenriched spectrum (top) and enriched spectrum (bottom). Citrullinated peptides identified before enrichment are marked by asterisks. (b) MALDI-TOF/TOF fragmentation spectrum of a CRB-enriched citrullinated peptide. Complete sequence coverage is observed. Source: Tutturen, 2010. . Reproduced with permission from Elsevier.
one depleted of IgG, and one where IgG and human serum albumin (HSA) were depleted. The synovial pellet was not fractionated. All four samples were digested with Lys-C and half of each digest enriched using the BPG method described earlier. Each sample was analyzed in triplicate; the samples were separated on a 50 cm C18 column over a 295 min gradient. Eluting peptides were fragmented with HCD. Analysis of the supernatant samples without purification or BPG enrichment resulted in confident identification of 119 citrul- linated peptides. Analysis of the unpurified supernatant with enrichment revealed 3673 unique monoisotopic masses from fragmentation spectra containing the m/z 270.13 BPG fragment ion (2146 unique masses from the IgG- and HSA-depleted sample). Only 4% of the peptides identified had fragmentation spectra missing the BPG ion. This suggests that the enrichment is very efficient. The database search results of the BPG enrichment samples
Figure 7.8 A schematic of the biotin-based BPG enrichment method. Peptides are mixed with BPG, and unreacted BPG is removed by SCX. The tagged peptides are enriched on streptavidin beads, and unmodified peptides washed off. The modified citrullinated peptides are eluted from the streptavidin and analyzed by MALDI-TOF. Source: Tutturen, 2013. . Reproduced with permission from Elsevier.
were however somewhat disappointing. Only 13 and 10 citrullinated peptides were identified from the supernatant and the depleted supernatant (IgG and HSA), respectively. It was noted that HCD fragmentation of peptides containing the BPG modification resulted in significant signal for m/z 270.13 along with m/z 227.1 and 286.1 (further backbone fragmentation of the BPG tag). The authors suggest that these signals reduced the fragment ion yield for peptide backbone fragments, resulting in poor sequence coverage and low-quality MS/MS spectra. Tutturen et al. acknowledge that there is room to improve the enrichment by modifying the BPG to contain a cleavable site. It is also possible that the use of ETD rather than HCD would not result in fragmentation of the BPG. The two enrichment protocols developed by Tutturen et al. to date have not been applied by other groups, but from the preliminary findings it is hopeful that large-scale citrullinomics is a possibility.