Biochemical Regulatory Features of Activation-Induced Cytidine Deaminase Remain Conserved from Lampreys to Humans

Abstract

Activation-induced cytidine deaminase (AID) is a genome-mutating enzyme that initiates class switch recombination and somatic hypermutation of antibodies in jawed vertebrates. We previously described the biochemical properties of human AID and found that it is an unusual enzyme in that it exhibits binding affinities for its substrate DNA and catalytic rates several orders of magnitude higher and lower, respectively, than a typical enzyme. Recently, we solved the functional structure of AID and demonstrated that these properties are due to nonspecific DNA binding on its surface, along with a catalytic pocket that predominantly assumes a closed conformation. Here we investigated the biochemical properties of AID from a sea lamprey, nurse shark, tetraodon, and coelacanth: representative species chosen because their lineages diverged at the earliest critical junctures in evolution of adaptive immunity. We found that these earliest-diverged AID orthologs are active cytidine deaminases that exhibit unique substrate specificities and thermosensitivities. Significant amino acid sequence divergence among these AID orthologs is predicted to manifest as notable structural differences. However, despite major differences in sequence specificities, thermosensitivities, and structural features, all orthologs share the unusually high DNA binding affinities and low catalytic rates. This absolute conservation is evidence for biological significance of these unique biochemical properties.

Keywords: biochemistry; enzymes; evolution; evolutionary immunology.

FIG 1

AID in the context of evolution. (A) Phylogenetic tree showing the evolutionary relationships between early-diverged fish species and humans: Petromyzon marinus (lamprey), Ginglymostoma cirratum (nurse shark), Tetraodon nigrovirigis (tetraodon), Latimeria chalumnae (coelacanth), and Homo sapiens (human). Images of the species were illustrated by Emma Quinlan. (B) Percent sequence identity and similarity were calculated by comparing identical and similar amino acids between the indicated AID orthologs and Hs-AID. To investigate the relative contribution of the ray-finned fish insert to sequence similarity, identical and similar homologies between Tn-AID and Hs-AID were calculated without this insert. The approximate period of species appearance (mya, million years ago) is shown on the x axis. Sequence identity and similarity to AID are based on published sequences of AID; some are predicted: 0 mya, Homo sapiens; 50 mya, Callithrix jacchus; 100 mya, Pteropus vampyrus; 150 mya, Ornithorhynchus anatinus; 300 mya, Anolis carolinensis; 350 mya, Pleurodeles waltl; 400 mya, Latimeria chalumnae and Tetraodon nigroviridis; 450 mya, Ginglymostoma cirratum; and 500 mya, Petromyzon marinus. (C) Alignment of AID protein sequences from human (Hs-AID), coelacanth (Lc-AID), tetraodon (Tn-AID), nurse shark (Gc-AID), and lamprey (Pm-CDA1). L, loop; α, helix; β, strand. (D) Percent sequence identity was calculated by comparing the nucleotide sequence between the AID-encoding cDNA sequences of fish AID orthologs and Hs-AID. The approximate period of species appearance is shown on the x axis. The cDNA sequences of the same nonfish species used to provide context for amino acid conservation were again used: 0 mya, H. sapiens; 50 mya, C. jacchus; 100 mya, P. vampyrus; 150 mya, O. anatinus; 300 mya, A. carolinensis; 350 mya, P. waltl; 400 mya, L. chalumnae and T. nigroviridis; 450 mya, G. cirratum; and 500 mya, P. marinus.

FIG 2

AID orthologs are active cytidine deaminase enzymes. (A, top) TGCbub7 substrate. *, ³²P label. The arrow shows the target cytidine in the bubble. (Bottom) Representative alkaline cleavage gel showing activity of 0.5 μg of Pm-CDA1, Gc-, Tn-, Lc-, and Hs-AID, and their corresponding catalytically dead mutants on TGCbub7, incubated at each enzyme's optimal temperature (14.5, 20.5, 20.0, 25, and 31°C, respectively). The substrate and product are 56 and 28 nucleotides in length, respectively. The negative control is the substrate TGCbub7 and activity buffer with no enzyme added. (B, left) Schematic of p219 showing the principle of the PCR assay for detecting AID activity on long stretches (several hundred nucleotides) of plasmid ssDNA. Outside primers: P1F (GGAAGGTATGAAAATAGGAAAAGAAAATAAATAAATTTTG) and P1R (CCCCTAACTTTTATACCCAACCCTAACTCC). Nested primers: P2F (CCCCCCGATCCGTATTTTTGGATAGTTAGGTGGT) and P2R (CCCCCCGATCCAATTAACAACCCTAAAATATAA). Reverse primers are designed to preferentially anneal with deaminated cytidines; the forward primers anneal to the replicated mutations. (Right) PCR amplification of a section of p219 plasmid untreated (AID neg) or incubated with five AID orthologs or GST at their optimal temperature (25°C for GST) twice for 1 h each. For PCRs to amplify AID-mutated sequences, p219 plasmid incubated with Hs-AID, Pm-CDA1, and GST was annealed to deamination-specific primers at 52°C; Lc-, Tn-, and Gc-AID reactions were annealed at 51°C.

FIG 3

Fish AID orthologs are cold adapted compared to Hs-AID. (A) Representative AID activity alkaline cleavage gels and data from thermosensitivity assays. Gels show AID incubated with TGCbub7 for temperatures ranging from 4 to 40°C. Graphs present thermosensitivity curves, the peak of which indicates the AID ortholog's optimal temperature. The y axis shows the percentage of maximum deamination. Error bars represent standard deviations (SDs). (B) Graph of average optimal temperatures for each AID ortholog, indicated by name in the x axis, with each organism illustrated and the final average optimal temperature above each bar. Each average is determined from three to six individual experiments using two to four independently purified preparations expressed from two or three independently constructed expression vector clones for each AID ortholog, resulting in, from left to right, 7, 4, 5, 12, and 5 independent experiments. Error bars represent the standard errors of the mean (SEM).

FIG 4

AID orthologs exhibit unique substrate specificity. (A) Representative alkaline cleavage gel of the substrate specificity assay, showing activity of Hs-AID on six bubble substrates (TGCbub7, TACbub7, AGCbub7, GACbub7, GTCbub7, and GGCbub7). (B) Bar graphs showing substrate specificity of each AID ortholog on six bubble substrates with the following target sequence: TGC, TAC, AGC, GGC, GTC, and GAC. Two or three purifications of one or two clones from each AID orthologs were tested two or three times for a total of three to five independent experiments for each species. The y axis shows relative deamination efficiency (preference relative to average of all six substrates) to enable comparison between AID orthologs whose absolute activity levels on each substrate vary. Error bars represent the SEM. *, P ≤ 0.05; **, P < 0.005. P values were determined by using a Mann-Whitney test. (C) Visualization of mutations. Each horizontal line is one individual sequenced amplicon of p219 that was mutated by the indicated AID ortholog and PCR amplified; each “X” is a C→T mutation. The x axis maps the section of p219 that is PCR amplified, minus the primers: 407 nucleotides from the first set of primers. The line at the top of each graph shows all possible dC mutations, indicated by an X. The numbers of amplicons analyzed for each AID ortholog were as follows: Pm-CDA1, 61; Gc-AID, 82; Tn-AID, 103; Lc-AID, 112; and Hs-AID, 114. (D) Comparison of substrate specificity of Hs-AID to early-diverged AID orthologs on p219. The y axis shows the mutability index, where 1 = average rate of mutation (dotted line) for all 16 NNC motifs. The relative preference for each individual NNC sequence was obtained by dividing its mutation rate by the average value for all 16 NNCs. The x axis shows XXC DNA motifs, ordered from most to least mutated by Hs-AID. Error bars represent the SD. (E) Bar graphs showing the substrate specificity of each AID ortholog on six NNC motifs—TGC, TAC, AGC, GGC, GTC, and GAC—in a PCR-based assay. The y axis shows the mutability index. The relative preferences for the six NNC sequences were obtained by dividing the mutation rate by the average value for the six NNC motifs identified. *, P ≤ 0.05; **, P < 0.005, ***, P < 0.0001. P values were determined by using a Mann-Whitney test. Error bars represent the SEM. (F) Sequence logos showing the relative AID specificity of each nucleotide at the −2 and −1 nucleotide positions. The height of the stack shows the consensus of nucleotides at that position, and the height of each symbol within the stack indicates the relative frequency of each nucleotide at that position.

FIG 5

Early-diverged AID orthologs exhibit lower enzymatic efficiency compared to Hs AID. (A, left) Representative time course alkaline cleavage gel. Pm-CDA1 was incubated with TGCbub7 at optimal temperature (∼14°C) for 1 to 20 h. (Right) Graph of combined time course data from two to seven experiments using three to six independently purified preparations of each AID ortholog; each point on the graph represents four to ten individual data points. Error bars represent the SD. (B, left) Representative Michaelis-Menten enzyme kinetics alkaline cleavage gel used to determine the initial enzymatic velocity of each AID ortholog. A 0.2-μg portion of Pm-CDA1 was incubated with concentrations of TGCbub7 ranging from 3.15 to 100 fmol/μl at the optimal temperature (∼14°C) for 7 h. (Right) Graph of Michaelis-Menten kinetics showing initial velocities of the different AID orthologs. Each point represents four to six individual data points from three independent experiments on three to six independently purified preparations of each AID ortholog. Error bars represent the SD. **, P < 0.005; ***, P = 0.0005 (determined by two-tailed, nonparametric t test).

FIG 6

Differences in activity between AID orthologs are not due to ssDNA binding. (A) Representative EMSA gel of Hs-AID incubated with various concentrations of TGCbub7 at optimal temperature (31°C) for 1 h. AID in complex with substrate is found in the bound bands, while free substrate continues to migrate toward the bottom of the gel, as indicated. The negative control is TGCbub7 in binding buffer with no AID added. (B) In order to calculate K_d half-saturation binding affinities for each ortholog, the fraction of shifted substrate was quantitated, and a bound versus free plot was generated. The average K_d value for each AID ortholog is indicated within each graph, determined from three independently purified preparations of each AID ortholog.

FIG 7

Predicted structures and substrate docking of AID orthologs. Shown are the cartoon (first column), surface topology (second column), surface model with single-stranded section of TGCbub7 in the DNA binding groove (third column), and catalytic pocket docked with dC (fourth column) of Hs-AID, Lc-AID, Tn-AID, Gc-AID and Pm-CDA1. Each structure is a representative of the 25 lowest-energy predicted conformations, using multiple related APOBEC X-ray and NMR structures as the templates. N-to-C terminus progression is shown from blue to red in the cartoon structures. The ribbon diagram shows Zn coordinated in the catalytic pocket (purple sphere): ℓ is loop, α is helix, and β is strand. All models illustrate the same global architecture with notable differences found in the N-terminal tails, connecting loops (ℓ2, ℓ4, ℓ5, ℓ6, and ℓ8), particularly in the loop 5 extension that is unique to bony fish, and in the structure/absence of the C-terminal α7. In the surface structures, positive and negative residues are blue and red, respectively. The catalytic pocket is seen as an indentation in the center, with the catalytic residues (Zn-coordinating triad of cysteines and histidine, catalytic glutamic acid) colored purple. A large proportion of the surface is positively charged, as reflected by the relatively high pI and net charge at neutral pH compared to other APOBEC enzymes, and similar to Hs-AID. The fourth column shows catalytically accessible conformations of each AID ortholog docked with dC in the catalytic pocket. All orthologs exhibited both open and closed catalytic pocket confirmations, similar to Hs-AID, and a representative open pocket confirmation was chosen for dC docking, since closed conformations are unable to dock dC in the pocket and thus represent an inactive conformation of AID.