Multiple myeloma (MM) is a hematologic cancer characterized by an abundance of bone marrow monoclonal plasma cells. Approximately half of patients with both MM and kidney disease have LCCN, characterized by multiple intraluminal light chain casts obstructing the renal tubules, which is a leading cause of acute kidney injury (AKI) [1], [2], [3]. The common features of pathogenic FLCs in cast formation still needs more investigation, but it can be speculated that their amino acid sequence is important. Determining exact sequences of patient-derived LCs to build a database is therefore highly desirable [4].

The sequence of FLCs can be partially determined by high-throughput sequencing assay from bone marrow mRNA encoding immunoglobulins (5′ rapid amplification of cDNA ends sequencing, 5′ RACE-seq); through this technique, full-length V(D)J region (variable, diversity and joining genes) and part of the C (constant gene) region of the monoclonal immunoglobulin can be determined [5]. However, bone marrow samples are difficult to obtain and of inconsistent, bringing difficulties to 5′ RACE-seq. Furthermore, this approach gives no insight into potential posttranslational modifications (PTMs) and full sequence of FLCs including the constant region, which may be important as well.

Recently, LC-MS/MS has been extensively used for antibody characterization. The predominant common approach is bottom-up proteomics (BUP), which relies on the protein digestion followed by LC-MS/MS analysis of the resulting peptides. BUP can provide high sequence coverage confirming the sequence of recombinant antibodies and the presence of expected PTMs in particular when using a combination of enzymes [6]. Unknown antibodies required de novo sequencing, which is much more difficult [7]. FLCs have a molecular weight in the range of 25 kDa, enabling the utilization of top-down proteomics (TDP) that involves the fragmentation of intact proteins [8]. Sequence gaps often remain because of the lack of fragment ions both in BUP and TDP; therefore, a combination of BUP and intact mass TDP with specific data analysis can be feasible [4], [9]. For instance, this approach has been used to classify plasma cell disorders (including amyloidosis) by analyzing of monoclonal immunoglobulin light chains in human serum [10], as well as determine abundantly produced FLCs in urine of MM patients [4].

Reducing the error rate and eliminate sequence gaps therefore become key issues. Instrumental errors, signal loss, and noise often mislead the de novo sequencing algorithms, resulting in a high error rate [11]. Although the combination of BUP and TDP can help overcome sequence gaps, the data analysis process can be is quite intricate and demanding. While BCR-seq has demonstrated to be effective and reliable enough to determine the sequence of pathogenic FLCs, a non-invasive and bone marrow independent method is highly desirable as well. To overcome these limitations, we developed a de novo sequencing workflow based on LC-MS/MS for protein sequencing of patient-derived FLCs.

In this workflow, only BUP approach was employed. The LC–MS/MS method was optimized to obtain the high-quality data required for de novo sequencing. Specifically, the resulting peptides by various enzymic digestion were analyzed through various fragmentation methods (HCD, ETD followed by EThcD) to maximize the coverage. Paired profiles generated by different fragmentation methods (e.g. ETD/HCD) can be used for de novo sequencing, thereby reconstructing a peptide sequence and optimizing both spectrograms [12]. The richer MS/MS spectra derived from EThcD(a combined fragmentation scheme by combining electron transfer and higher-energy collision dissociation) shows advantages for peptide de novo sequencing and identification of PTMs.

PEAKS can perform all of the above types of analysis and was used for the de novo sequencing of peptides. Full-length FLCs were further assembled and compared with the corresponding results from 5′RACE-seq. The reliability of de novo protein sequencing can be validated by comparing whether the two methods yield the same results. Moreover, the characteristics of these two pathogenic FLC were analyzed, this may give some new information of pathogenic mechanisms.