A stunningly elegant mechanism underlies the adaptive immune system's ability to create a wide assortment of T-cell receptors (TCRs). Each TCR is a complex of membrane proteins unique to a T-cell clone; collectively, TCRs allow the recognition, recall, and destruction of the many pathogenic agents that vertebrates encounter. The vast majority of TCRs are heterodimers composed of two distinct subunit chains (α and β), both containing variable domains. The term "clonotype" is used to refer either to a particular TCR variant (TCR-α or TCR-β subunit) or to a particular pairing of TCR subunit variants (TCR-α and TCR-β) shared amongst a clonal population of T cells. TCR diversity is generated during the early stages of T-cell development when extensive recombination occurs between the V- and J-segments, and the V-, D-, and J-segments, in the TCR-α and TCR-β genes, respectively, via a mechanism that also incorporates and deletes additional nucleotides. This process-commonly referred to as "V(D)J recombination"-as well as a selection process that eliminates self-antigen-recognizing T cells, yields a T-cell repertoire with sufficient TCR diversity to collectively recognize any antigen.

The regions of the TCR-α and TCR-β genes that encode the "complementarity determining region 3" (CDR3), which makes primary contact with the antigen-MHC complex, are thought to be unique to each TCR-α and TCR-β variant; therefore, sequence variation in the CDR3 region serves as a useful proxy for overall T-cell repertoire diversity and is frequently used to quantify diversity in high-throughput profiling experiments. TCR-profiling experiments typically focus on capturing genomic DNA or mRNA sequences that correspond to the CDR3 region. While sequencing genomic DNA may be preferable for certain TCR-profiling applications-including those that involve quantifying various T-cell subpopulations-this approach is not without its limitations, and methods that involve analyzing mRNA sequences carry several important advantages:

• TCR mRNA templates are likely to be more highly represented than DNA templates in any one T cell, such that mRNA sequencing approaches will afford greater sensitivity and allow for more comprehensive identification of unique TCR variants, including those that are present in a very small proportion of T cells.
• Sequencing mRNA rather than genomic DNA specifically allows for the identification of expressed TCR sequences that have undergone splicing and post-transcriptional processing and are more likely to yield functional proteins. DNA-based approaches, by contrast, do not identify TCR sequences in their translated forms, and will unavoidably yield many functionally irrelevant sequences.

Our high-throughput method for TCR mRNA profiling (Figure 1) leverages SMART (Switching Mechanism at the 5′ end of RNA Template) technology, a 5'-RACE-based approach that enables efficient capture of full-length transcripts. This is followed by semi-nested PCR to fully capture and amplify TCR-α and TCR-β variable regions and prepare libraries for sequencing on Illumina® platforms. A major advantage of this method is the avoidance of multiplex PCR, which increases the likelihood of sample misrepresentation due to amplification biases.

Figure 1. Library preparation workflow and PCR strategy for TCR profiling using the SMARTer approach. Panel A. Reverse transcription and PCR amplification of TCR subunit mRNA sequences. First-strand cDNA synthesis is primed by the TCR dT Primer and performed by an MMLV-derived reverse transcriptase (RT). Upon reaching the 5′ end of each mRNA molecule, the RT adds non-templated nucleotides to the first-strand cDNA. The SMART-Seq v4 Oligonucleotide contains a sequence that is complementary to the non-templated nucleotides added by the RT, and the oligo hybridizes to the first-strand cDNA. In the template-switching step, the RT uses the remainder of the SMART-Seq v4 Oligonucleotide as a template for the incorporation of an additional sequence on the end of the first-strand cDNA. Full-length variable regions of TCR cDNA are selectively amplified by PCR using primers that are complementary to the oligonucleotide-templated sequence (SMART Primer 1) and the constant region(s) of TCR-α and/or TCR-β subunits (TCRa and/or TCRb Mouse Primer 1). A subsequent round of PCR is performed to further amplify variable regions of TCR-α and/or TCR-β subunits and incorporate adapter sequences using a TCR Primer 2 Forward HT Index and a TCRa and/or TCRb Mouse Primer 2 Reverse HT Index. Included in the primers are adapter and index sequences (Read 2 + i7 + P7 and Read 1 + i5 + P5, respectively) that are compatible with the Illumina sequencing platform. Following purification, size selection, and quality analysis, the TCR cDNA library is ready for sequencing. Panel B. Semi-nested PCR approach for amplification of TCR-α and/or TCR-β subunits. The primer pairs used for the first round of amplification capture the entire variable region(s) and most of the constant region(s) of TCR-α and/or TCR-β cDNA. The second round of amplification retains the entire variable region(s) of TCR-α and/or TCR-β cDNA and a smaller portion of the constant region(s). The anticipated size of the final TCR library cDNA (inserts + adapters) is ~700-800 bp.

The SMARTer Mouse TCR a/b Profiling Kit (Cat. Nos. 634402, 634403 & 634404) allows unparalleled sensitivity in the analysis of TCR diversity, including detection of low-abundance TCR variants, from 10 ng-500 ng of total RNA obtained from mouse immune tissues/cells including spleen, thymus, and PBMCs, or from 1,000-10,000 purified T cells. For comprehensive background information on TCR profiling and for data on the sensitivity and reproducibility of our TCR a/b profiling kits, please refer to the SMARTer Human TCR a/b Profiling Kit technical note. The mouse-specific kit uses a similar method with mouse sequence-specific primers; although the target species differs from the human kit, results from the human kit are a testament to the power of the overall method.

In order to assess the performance of the mouse-specific kit, we collaborated with two investigators, who used our technology to answer the following questions:

• The first question addressed accuracy and coverage: 
  What read depth is required for a confident estimation of clonotype diversity?
• The second question addressed sensitivity of detection: 
  How well can the method identify low-abundance clonotypes?

Elucidating Clonotype Diversity

The goal of our initial collaboration with investigators Dr. Karine Bourgade and Prof. Joost van Meerwijk at Inserm in Toulouse, France was to examine the diversity of the clonotypes identified in mouse spleen. Mouse splenocytes and CD4+ T cells were isolated from four mouse spleens as described in the Methods section, and total RNA was purified from each population (Figure 2A). Libraries were prepared using the SMARTer Mouse TCR a/b Profiling Kit, and samples were sequenced and analyzed as described in the Methods section. All samples showed a good on-target mapping rate: approximately 70% of total reads aligned to genes encoding TCR-α or TCR-β, with approximately equal proportions of reads aligning to each gene (Figure 2B).To determine the read depth required to accurately estimate TCR diversity, our collaborators investigated the number of reads at which saturation in the number of unique TCR-α or TCR-β clonotypes occurs (Figure 2C). With the splenocyte samples, saturation of clonotype identification occurred at ~100,000 reads for both TCR-α and TCR-β. However, as expected, the enriched CD4+ T-cell populations showed greater diversity such that 400,000 reads were required to approach saturation for both TCR-α and TCR-β; above 400,000 reads, the rate of clonotype identification decreased as saturation was approached.

Figure 2. Assessment of TCR diversity in mouse spleen samples with the SMARTer Mouse TCR a/b Profiling Kit. Panel A. To compare TCR diversity between the spleen's general T-cell population and purified CD4+ T cells, total splenocytes were isolated from spleen and divided into two groups. The first group was left untouched, and the second group was subject to CD4+ T-cell purification (see Methods section for details). Total RNA was extracted from the total splenocyte group and the group enriched for CD4+ T cells, and each RNA sample was used for library preparation. Panel B. Proportion of on-target reads mapping to TCR-α or TCR-β CDR3 sequences. Panel C. Rarefaction curves showing the relationship between read depth and the number of unique clonotypes identified for TCR-α (left) and TCR-β (right). When a curve starts to flatten off, the number of unique clonotypes identified is reaching saturation. Dashed lines show the theoretical number of clonotypes as the number of reads increases.

Taken together, these data show that the SMARTer Mouse TCR a/b Profiling Kit enables on-target mapping for both TCR-α and TCR-β subunits, and consequently, the reproducible grouping of cell populations into more diverse (CD4+ T-cell) and less diverse (splenocyte) subsets, as expected. While these data have allowed us to determine the read depth required to fully capture TCR diversity for different sample types at a 10 ng input amount, it is likely that more reads would be required to achieve saturation of clonotype identification at greater input amounts.

Identifying Low-Abundance TCR Clonotypes

Our second question, how well can our method identify low-abundance clonotypes, was addressed in collaboration with Dr. Helen McGuire and Prof. Barbara Fazekas de St. Groth at the Centenary Institute in Sydney, Australia.Previous studies established that Rag-1-deficient (Rag-1^-/-) transgenic (Tg) mice expressing co-integrated, rearranged TCR-α and TCR-β chains spontaneously develop colitis with increasing age (Koh et al. 1999). Given that V(D)J recombination and T-cell differentiation are disrupted in Rag-1^-/-mice, all CD4+ T cells in the Tg animals presumably express the rearranged, transgenic TCR subunits. In the initial stages of the current study, flow cytometric analysis of CD4+ T cells isolated from Tg mice aged 16, 34, or 38 weeks identified the cells that expressed the Tg-encoded TCR-β chain but that lacked expression of the Tg-encoded TCR-α chain. These rare cells were present at higher frequencies in gut tissues and gut-draining lymph nodes. To investigate the nature of these rare cells further and to identify the nondominant TCR-α clonotypes present, the SMARTer Mouse TCR a/b Profiling Kit was used to generate sequencing libraries from total RNA extracted from various tissues (thymus, spleen, peripheral lymph nodes, mesenteric lymph nodes, and large intestine) of the Tg mice (aged 16, 34, and 38 weeks). As TCR-α was of greater interest than TCR-β in this study, primers for the gene-specific PCRs were mixed at a ratio of 2:1 (TCRα:TCRβ), rather than the recommended 1:1 ratio outlined in the kit protocol. Skewing the primer ratio in this manner allowed for greater representation of TCR-α sequences in the resulting libraries and sequencing data, while still yielding sequence information for the beta chain. The TCR sequences identified were attributed to CD4+ T cells, given that CD8+ T cells do not develop in this particular Tg Rag-1^-/- strain. Following sequencing, read files were downsampled to 400,000 reads per sample based on the read saturation points identified previously (Figure 2). Analysis was performed using MiXCR (Bolotin et al. 2015) with the TRA locus specified. (All other parameters were set as default.) ~60-80% of reads mapped to TCR-α CDR3 sequences (Figure 3A), indicating that the skewing of the primer ratio successfully yielded more sequencing reads mapping to TCR-α than TCR-β sequences.

Figure 3. Assessment of TCR-α clonotype diversity in various tissues of transgenic mice of different ages with the SMARTer Mouse TCR a/b Profiling Kit. Panel A. The percentage of reads mapping to TCR-α CDR3 sequences identified in the thymus, spleen, peripheral lymph nodes (PLN), mesenteric lymph nodes (MLN), and large intestine (LI) is consistent across the 16-week-old, 34-week-old, and 38-week-old mice. Panel B. The immune tissues showed the highest TCR-α clonotypic diversity, while the large intestine showed the lowest diversity. This indicates that the large intestine has the lowest T-cell infiltrate. Panel C. Abundance of the TRAV3-2-TRAJ58 clonotype, expressed as a percentage of all reads mapping to any TCR-α or TCR-β clonotype, across tissue types and ages. Panel D. Abundance of the TRAV16N-TRAJ56 clonotype, expressed as a percentage of all reads mapping to any TCR-α or TCR-β clonotype, across tissue types in the 38-week-old mouse.   

To examine the relative diversity of TCR-α, the number of unique TCR-α clonotypes was identified for each of the aforementioned tissues from each mouse (Figure 3B). High clonotypic diversity was seen for the samples derived from immune tissues, but diversity was dramatically decreased in tissue of the large intestine, presumably due to low T-cell infiltration and/or expansion of specific clonotypes in response to gut-derived antigens.As expected, for all mice tested, the most highly represented clonotypes included the dominant Tg-encoded V gene segments TRAV4-4 (TCR-α) and TRBV26 (TCR-β). Subsequent analysis focused on identification of clonotypes containing nondominant TCR-α V gene segments. TRAV3-2-TRAJ58 was identified in all samples as a minor fraction of the total clonotype reads; in the thymuses of the 16-week-old and 34-week-old mice, this clonotype made up ~0.025% of all clonotypes, whereas in the 38-week-old mouse, its abundance was <0.005% (Figure 3C). For all mice tested, the abundance of this clonotype was highest in the mesenteric lymph node tissue and the lowest in the large intestine. Private clonotypes (those rarely observed in multiple individuals) showing tissue-specific or age-specific differences were also examined. The TRAV16N-TRAJ56 clonotype was present in all tissue samples from the 38-week-old mouse. Intriguingly, the abundance of this clonotype was the greatest in the large intestine-the site of inflammation in this mouse model-suggesting that elevation in the frequency of T cells carrying this TCR rearrangement (due to gut-antigen exposure) may be involved in the development of colitis (Figure 3D).

The SMARTer Mouse TCR a/b Profiling Kit provides a powerful solution for those seeking an accurate and sensitive approach to the analysis of mouse TCR diversity. In contrast with profiling methods that involve the amplification of genomic DNA, our 5'-RACE-based approach uses total RNA as input material. Starting from RNA allows for the capture of complete TCR V(D)J variable regions, the avoidance of multiplex PCR amplification strategies, and the detection of low-abundance TCR clonotypes. Independent groups confirmed that the kit enables a confident estimation of clonotype diversity-including identification of extremely low-abundance clonotypes-and provides a complete and accurate description of TCR repertoires from a range of mouse sample types and input amounts.

All sequencing libraries were generated using protocols and reagents included in the SMARTer Mouse TCR a/b Profiling Kit (Takara Bio, Cat. Nos. 634402, 634403 & 634404).

Input material

Libraries generated for the first study (Figure 2) used 10 ng of total RNA from an unsorted splenocyte population and 10 ng of total RNA from CD4+ T cells purified from total spleen cells using the Dynabeads Untouched Mouse CD4 Cells Kit (Life Technologies). In the second study (Figure 3), libraries were generated from 10 ng RNA isolated from the thymus, spleen, peripheral lymph nodes, mesenteric lymph nodes, or large intestine of 16-week-old, 34-week-old, or 38-week-old 5C.C7 Rag-1^-/- TCR Tg mice (-D line).

PCR parameters

As indicated in the SMARTer Mouse TCR a/b Profiling Kit User Manual, all libraries were subject to two rounds of semi-nested PCR amplification. In both of the studies shown here, all libraries were generated from 10 ng input RNA; 21 cycles of PCR1 and 20 cycles of PCR2 were performed. For the second study (Figure 3), a 2:1 (TCRα:TCRβ) ratio was used when mixing gene-specific primer sets in order to achieve a greater representation of TCR-α sequences.

Library purification and validation

Amplified libraries were purified using the Agencourt AMPure XP PCR purification kit (5-ml size: Beckman Coulter, Item No. A63880; 60-ml size: Beckman Coulter, Item No. A63881). Libraries were subject to one double size selection as per the user manual; beads were pelleted using a Magnetic Separator - PCR Strip (Takara Bio, Cat. No. 635011). Library validation was performed on an Agilent 2100 Bioanalyzer with the DNA 1000 Kit (Agilent, Cat. No. 5067-1504), using 1 µl of purified, undiluted library.

Illumina sequencing

Libraries were pooled to a final pool concentration of 4 nM. Pooled libraries were then diluted to a final concentration of 13.5 pM, including a 10% PhiX Control v3 (Illumina, Cat. No. FC-110-3001) spike-in. Libraries were sequenced on an Illumina MiSeq® sequencer using the 600-cycle MiSeq Reagent Kit v3 (Illumina, Cat. No. MS-102-3003) with paired-end, 2 x 300 base pair reads.

Data analysis

In the first study (Figure 2), sequencing data was analyzed using the MiXCR software package (Bolotin et al. 2015) and VDJTools (Shugay et al. 2015). In the second study (Figure 3), sequencing data was downsampled to 400,000 reads per sample and data analysis was performed using MiXCR.

Bolotin, D. A. et al. MiXCR: software for comprehensive adaptive immunity profiling. Nat. Methods 12, 380-1 (2015).

Koh, W. P. et al. TCR-mediated involvement of CD4+ transgenic T cells in spontaneous inflammatory bowel disease in lymphopenic mice. J. Immunol. 162, 7208-16 (1999).

Shugay, M. et al. VDJtools: Unifying Post-analysis of T Cell Receptor Repertoires. PLoS Comput. Biol. 11, e1004503 (2015).