Site-specific chemical modification of recombinant proteins produced in mammalian cells by using the genetically encoded aldehyde tag

Abstract

The properties of therapeutic proteins can be enhanced by chemical modification. Methods for site-specific protein conjugation are critical to such efforts. Here, we demonstrate that recombinant proteins expressed in mammalian cells can be site-specifically modified by using a genetically encoded aldehyde tag. We introduced the peptide sequence recognized by the endoplasmic reticulum (ER)-resident formylglycine generating enzyme (FGE), which can be as short as 6 residues, into heterologous proteins expressed in mammalian cells. Cotranslational modification of the proteins by FGE produced products bearing a unique aldehyde group. Proteins bearing this “aldehyde tag” were chemically modified by selective reaction with hydrazide- or aminooxy-functionalized reagents. We applied the technique to site-specific modification of monoclonal antibodies, the fastest growing class of biopharmaceuticals, as well as membrane-associated and cytosolic proteins expressed in mammalian cells.

Recombinant proteins are now among the portfolio of clinical candidates in most major pharmaceutical companies (1). Examples include growth factors, hormones, cytokines, replacement enzymes, clotting factors, and monoclonal antibodies, the fastest growing class of protein drugs (2). In many cases, the intrinsic properties of therapeutic proteins can be improved by chemical modifications (3). The most familiar of these is covalent attachment of polyethylene glycol (PEG) chains, a process also termed PEGylation, which can improve pharmacokinetic properties (4). Chemical modification can also expand a protein's capabilities, such as conversion of an antibody into a drug or diagnostic targeting element (5).

The major technical challenge in protein modification is achieving site selectivity for the production of a homogenous product. Most protein modification methods exploit the reactivities of endogenous functionality such as the amino group of lysine and the thiol group of cysteine (6). However, because proteins often possess multiple copies of these residues, site-specific labeling can be difficult to achieve. Thus, several groups have sought to introduce unique functionalities into proteins that are chemically orthogonal to the 20 proteogenic amino acid side chains. One approach involves engineering unique combinations of natural residues such as the tetra cysteine motif that reacts with biarsenical probes (7). Alternatively, a unique functional group that is not found in natural amino acids can be installed at the intended modification site and then reacted in a second step with the moiety of interest. Ketones and azides epitomize the kinds of functional groups that have been exploited for this purpose (8). These electrophiles can be site-specifically incorporated into proteins by using a variety of elegant methods, including chemical modification of a protein's N terminus (9), unnatural amino acid mutagenesis (10), or through the use of enzymes that transfer prosthetic groups to the protein (11–13). These methods are powerful research tools, but practical issues, such as scalability and generality across proteins and expression systems, may limit their use in pharmaceutical protein modification.

We recently reported a method for redirecting endogenous cellular machinery to introduce aldehydes into recombinant proteins. The method exploits formylglycine-generating enzyme (FGE) (14), which converts cysteine to formylglycine (FGly) within a conserved 13-residue consensus sequence found in Type 1 sulfatases (Fig. 1A) (15). The modification is thought to occur cotranslationally, before protein folding, and is critical for the sulfatases' catalytic function. We found that the consensus sequence can be installed within heterologous proteins expressed in Escherichia coli, where it is modified efficiently by a coexpressed bacterial FGE (16). Furthermore, the minimized 6-residue sequence LCxPxR, derived from the most highly conserved portion of the FGE recognition site, also directed efficient conversion of cysteine to FGly. The isolated proteins were thus outfitted with an aldehyde group for site-specific chemical modification with aminooxy- or hydrazide-functionalized moieties, including fluorophores, affinity tags, and PEG chains (Fig. 1B). We demonstrated the generality of the technique, which we call the “aldehyde tag,” in prokaryotic expression systems by modifying several proteins at different locations within their primary sequence.

Fig. 1.

Site-pecific protein modification using the genetically encoded aldehyde tag. (A) FGE oxidizes a critical cystein to FGly within a conserved 13-amino acid sequence. (B) The aldehyde tag can be transported into a heterologous protein for site-specific modification with chemical probes.

A vast majority of pharmaceutically relevant proteins, however, must be expressed in mammalian cells due to their requirement of various posttranslational modifications (17). Disulfide bonds, glycosylation, and sulfation, for example, are difficult or impossible to achieve in bacteria or yeast. Thus, we sought to extend the aldehyde tag technology to proteins expressed in mammalian cells. Conveniently, in eukaryotes, FGE is located in the endoplasmic reticulum (ER) where it modifies sulfatases destined for lysosomes or secretion (14). We proposed that heterologous proteins endowed with an aldehyde tag sequence might also be modified by mammalian FGE as they traverse the secretory pathway.

Here, we demonstrate that recombinant proteins expressed in mammalian cells can be site-specifically modified by using the aldehyde tag technology. We initially focused on monoclonal antibodies because of their clinical importance. The most widely used isotype, IgG, has a conserved N-linked glycosylation site in the Fc region of each heavy chain that is important for protein structure and for various effector functions (18). Because of this posttranslational modification as well as several disulfide bonds, functional IgGs can only be produced in mammalian (or insect) expression systems. Conventional methods for IgG labeling involve nonspecific coupling reactions with lysine residues on the protein surface. Because lysines may lie near the antigen-binding site of some antibodies, there is a risk that these labeling strategies will reduce antibody function. The aldehyde tag would offer a more controlled method for covalent modification, one that could potentially be tailored for any chosen site on the antibody structure.

Accordingly, we engineered both Fc and intact IgG constructs bearing aldehyde tags at various sites. When expressed in CHO or HEK cells, the proteins were efficiently modified with aldehyde groups that were exploited for site-specific modifications. Furthermore, we demonstrated that membrane-associated proteins can be labeled with aldehyde tags and then chemically derivatized on live cells. Finally, we found that a cytosolic protein expressed in HEK cells can be modified by a coexpressed prokaryotic FGE directed to that compartment. The aldehyde tag is therefore a robust and general method for modification of proteins produced in mammalian systems.

Results and Discussion

Expression and Chemical Modification of Aldehyde-Tagged Fc Domain.

Implementation of the aldehyde tag in mammalian systems was first explored by using the IgG Fc fragment as a model protein. This stable fragment expresses at high levels in many mammalian cells lines, affording the quantities needed for detailed chemical analysis. In addition, the Fc domain has therapeutic relevance on its own (19) and as a fusion with other proteins (20). Using the commercial pFuse vector (InvivoGen), we generated Fc constructs encoding either a 13-residue aldehyde tag sequence derived from human arylsulfatase A (LCTPSRAALLTGR, “Ald₁₃”) or the minimized 6-residue tag LCTPSR (“Ald₆”) situated at either the N or C terminus (4 constructs in total). As controls, we also generated constructs in which the cysteine residue modified by FGE was mutated to an alanine residue.

The constructs were introduced into Chinese hamster ovary (CHO) cells by transient transfection in serum-free medium and the expressed proteins were isolated by using protein A/G (a genetically engineered protein that contains Fc-binding domains of both protein A and protein G) agarose. All 4 proteins expressed at approximately the same levels, which were comparable with the expression level of the unmodified Fc fragment (≈1.0 mg/L). To probe for the presence of aldehydes, the proteins were reacted with an aminooxy-FLAG peptide (DYKDDDDK) conjugate (16) (Fig. 2A illustrates a C-terminally modified Fc fragment) at pH 5.5 and analyzed by 4–12% SDS/PAGE and Western blot. The Fc proteins bearing Ald₁₃ or Ald₆ tags at either the N or C terminus all showed robust labeling, whereas the control C to A proteins gave no detectable signal (Fig. 2B; only Fc proteins with C-terminal Ald₁₃ and N-terminal Ald₆ are shown; the other 2 proteins gave similar results). Tryptic digestion of the aldehyde-tagged Fc proteins allowed direct identification of FGly by mass spectrometry (Fig. 2C). From all 4 proteins, the FGly-containing tryptic peptide as well as the unconverted cysteine-containing peptide were readily identified.

Fig. 2.

Site-specific labeling of aldehyde-tagged IgG Fc domain. (A) Nucleotides encoding the 13-amino acid or 6-amino acid sequence can be appended to the gene of IgG Fc fragment. Upon expression, the encoded cysteine is modified to an aldehyde, which can be used as a chemical handle for reaction with an aminooxy FLAG probe. (B) (Upper) Western blot probed with α-FLAG antibody. (Lower) Coomassie blue-stained image of the gel. (C) Mass spectra confirming the presence of FGly in a tryptic peptide from C-Ald₁₃-Fc [(M + 2)/2]. (Left) Mass spectrum of the tryptic fragment incorporating FGly. Theoretical: 508.7613 m/z, observed: 508.7755. (Right) Mass spectrum of the tryptic fragment incorporating unmodified Cys after treatment with 2-iodoacetamide. Theoretical: 546.2721 m/z, observed: 546.2811.

To determine the extent of conversion of Cys to FGly for all of the Fc constructs, we doped an isotope-labeled peptide into the digestion products from individual Fc constructs as an internal standard for mass spectrometry quantitation (21). The internal standard possessed the same sequence as the Cys-containing peptide, thus allowing direct determination of its concentration within each Fc protein digest. By subtraction, the fraction of FGly-containing peptide could be calculated to determine the conversion efficiency. The overall conversion efficiency ranged from 25% to 67% for the various constructs tested (Table 1).

Table 1.

Percent conversion of Cys to FGly in recombinant Fc fragments

Tag location	Tag type	CHO cells	CHO cells transiently cotransfected with hFGE¶	CHO cells stably transfected with hFGE‖
N terminus*	6-mer‡	40 ± 8	44 ± 2	68 ± 12
N terminus	13-mer§	67 ± 1	91 ± 2	77 ± 3
C terminus†	6-mer	28 ± 1	45 ± 1	62 ± 3
C terminus	13-mer	45 ± 2	69 ± 3	64 ± 2

The standard deviation represents the error of at least 3 replicate experiments.

*The tag was inserted downstream of the signal peptide sequence.

^†The tag comprised the C-terminal residues of the protein. Two residues upstream of the tag sequence were mutated in order to generate a Kpnl restriction site.

^‡The 6-mer sequence was LCTPSR.

^§The 13-mer sequence was LCTPSRAALLTGR.

^¶Human FGE (hFGE) was encoded on a separate plasmid, pcDNA3.1.

^‖The procedure used for generating the stably transfected CHO cell line is in SI Text.

The relatively low conversion efficiencies for some of the constructs might reflect an inadequate supply of endogenous FGE for complete modification of the overexpressed Fc proteins as they mature in the ER. In prokaryotic systems, we found that co-overexpression of a heterologous FGE can increase the conversion efficiency of a tagged protein (16). Accordingly, we overexpressed human FGE (hFGE) in CHO cells, either transiently or stably, along with transient expression of the Fc constructs. A shown in Table 1, hFGE coexpression increased the conversion efficiency for each Fc construct. In most cases, the most dramatic improvements were observed with transiently overexpressed hFGE. For a list of the oligonucleotides used in this study, see Table S1.

The selectivity of the aldehyde-aminooxy condensation reaction was demonstrated by selective modification of aldehyde-tagged Fc protein in the context of crude conditioned medium. CHO cells were transiently transfected with N-Ald₁₃-Fc alongside hFGE. The conditioned medium was collected and dialyzed against PBS to remove competing metabolites (e.g., pyruvate, which has a reactive ketone, and glucose, which possesses an aldehyde) while retaining proteins. The dialyzed medium was reacted with aminooxy-FLAG and analyzed by Western blot. As shown in Fig. 3, the Fc protein was the only species detected, despite the presence of other abundant proteins in the sample. Notably, the only endogenous cellular proteins predicted to possess FGly residues are the lysosomal and secreted sulfatases as well as a few kinesin family member proteins (Table S2), none of which were visible in the CHO cell conditioned medium. In model studies, we observed no covalent labeling of commercial sulfatases with hydrazide probes, presumably because their FGly residues are buried in the active site of the folded protein (data not shown) (15). Thus, among a complex mixture of cellular proteins that includes active sulfatases, aldehyde-tagged heterologous proteins may be the only substrates that label with aminooxy- and hydrazide-functionalized reagents.

Fig. 3.

Site-specific labeling of N-Ald₁₃-Fc domain in crude conditioned medium. (A) Coomassie blue-stained image of the gel. (B) Western blot probed with α-FLAG antibody.

Modification of an Aldehyde-Tagged Full-Length IgG.

We extended the aldehyde tag method to site-specific modification of a full-length monoclonal IgG specific for Nogo receptor-2 (NgR2) (22). This antibody comprises Fab fragments derived from phage display, fused to a human Fc domain by using the pIgG construct (23). We cloned the 6-mer aldehyde tag sequence (LCTPSR) into the C terminus of the CH3 region of the heavy chain. This protein and the light chain, encoded in the same plasmid, were expressed transiently in HEK293T suspension cells alongside hFGE (2:1 wt/wt). The protein was purified from the conditioned medium by using protein A-agarose, then reacted with aminooxy-biotin and analyzed by SDS/PAGE and Western blot (Fig. 4A). The intact antibody was specifically labeled, and analysis of tryptic fragments by mass spectrometry confirmed the location of the modified aldehyde tag on the heavy chain (Fig. S2–S3). As before, a C108A mutation in the aldehyde tag sequence abrogated labeling. In addition, we conjugated the aldehyde-tagged antibody to aminooxy-Alexa Fluor 488 (Invitrogen) and investigated the ability of the labeled antibody to bind recombinant Nogo R2 expressed on HEK293 cells (Fig. 4B). Importantly, the modified antibody retained its antigen binding activity. In a control experiment, only background signals were detected when HEK293 cells expressing Nogo R1 were treated with the labeled antibody and analyzed by flow cytometry (data not shown).

Fig. 4.

Labeling of the aldehyde-tagged full-length IgG. (A) Site-specific labeling of the monoclonal α-Nogo R2 IgG with aminooxy biotin probe. (Upper) Western blot probed with α-biotin antibody. (Lower) Coomassie blue-stained image of the gel. (B) Fluorophore conjugated α-Nogo R2 IgG retains binding specificity.

Labeling of Aldehyde-Tagged Cell Surface Proteins.

Like secreted proteins, plasma membrane-associated proteins traffic through the secretory pathway and are therefore potential substrates for ER-resident FGE. To test this notion, we introduced the 13-mer aldehyde tag at the N terminus of the platelet-derived growth factor receptor (PDGFR) transmembrane (TM) domain encoded in the pDisplay vector (Invitrogen). The construct was transiently expressed in CHO cells along with hFGE. After 48 h, the cells were reacted with biotin hydrazide then stained with Alex Fluor 488-streptavidin and analyzed by flow cytometry (Fig. 5A). As shown in Fig. 5A, CHO cells expressing aldehyde-tagged PDGFR-TM (Ald₁₃-TM) showed a marked increase in fluorescence compared with cells expressing the unmodified PDGFR-TM protein or the C18A mutant. These results were mirrored by using fluorescence microscopy (Fig. 5B). We were not able to perform a detailed quantitation of Cys to FGly conversion because of the small quantities and complex structure of this membrane-associated protein.

Fig. 5.

Labeling of membrane-associated Ald₁₃-TM and Ald₁₃-CD4. (A) Flow cytometry analysis of CHO cells transfected with Ald₁₃-TM or unmodified PDGFR-TM (TM) or Ald₁₃-TM (C18A). At 48 h after transfection, cells were labeled with biotin hydrazide [1 mM in PBS (1% FBS) (pH 6.5) for 1 h] and stained with Alex Fluor 488-labeled streptavidin. Error bars represent the standard deviation of the mean for 3 replicate ligation reactions. Similar results were obtained in 3 replicate experiments. MFI, mean fluorescence intensity; AU, arbitrary units. (B) Fluorescence micrographs of CHO cells transfected with Ald₁₃-TM (Left) or TM (Right) and labeled with aminooxy Alexa Fluor 647. (C) Fluorescence micrographs of CHO cells transfected with Ald₁₃-CD4 (Left) or Ald₁₃-CD4(C18A) (Right) and labeled with Alexa Fluor 488. Dye labeling reactions were performed at pH 6.4 for 1 h at room temperature. Excess reagent was washed off, and the cells' nuclei were stained with Hoechst 33342 before imaging. Blue, DAPI channel; green, FITC channel; red, Cy5 channel.

Similarly, we inserted an aldehyde tag at the N terminus of recombinant mouse CD4, a membrane-associated protein involved in T cell activation (24). As demonstrated by using fluorescence microscopy, CHO cells expressing aldehyde-tagged CD4 (Ald₁₃-CD4) were specifically labeled with aminooxy-Alexa Fluor 488, whereas cells expressing the corresponding C18A mutant had only low levels of background fluorescence (Fig. 5C).

Labeling of an Aldehyde-Tagged Cytosolic Protein.

We were curious whether cytosolic proteins in mammalian cells could be labeled by using the aldehyde tag method. Endogenous FGE is reportedly confined to the ER, where it is glycosylated and possibly associated with other ER-resident proteins (14). Thus, for soluble cytosolic expression, we cloned a bacterial FGE homolog derived from Streptomyces coelicolor (25) into a mammalian expression vector. The GFP derived from Aequorea coerulescens was modified with the 13-mer aldehyde tag downstream of an N-terminal His₆ tag. This construct, called Ald₁₃-GFP, was transiently expressed in HEK293T cells along with the S. coelicolor FGE. After cell lysis and nondenaturing Ni-NTA purification, the protein was reacted with biotin-hydrazide and analyzed by nonreducing PAGE and Western blot.

As shown in Fig. 6, the aldehyde-tagged GFP, but not the C10A mutant, was effectively labeled with biotin hydrazide. Labeling depended on coexpression of S. coelicolor FGE. However, GFP expressed in the absence of S. coelicolor FGE appeared to undergo conversion of Cys to FGly at a low level. The Coomassie-stained gel of GFP expressed together with S. coelicolor FGE revealed 2 bands, one migrating at 32 kDa corresponding to the monomeric protein and another at 64 kDa corresponding to the disulfide-bound dimer. The ratio of these 2 species reflects the percent conversion of Cys to FGly. Accordingly, the C10A mutant appeared as a single band at 32 kDa. When GFP was expressed without S. coelicolor FGE, a majority of the protein migrated at 64 kDa, but a minor amount migrated with an apparent molecular mass of 32 kDa, consistent with a low level of Cys modification.

Fig. 6.

Labeling of Ald₁₃-GFP after cytosolic expression in HEK cells. The Ald₁₃-GFP and Ald₁₃-GFP (C10A) plasmids were transiently transfected into HEK 293T cells with (+) or without (−) S. coelicolor FGE. Three days after transfection, cells were lysed by grinding on ice, and proteins were purified on Ni-NTA agarose. (Upper) α-Biotin Western blot of aldehyde-tagged GFP constructs after reaction with biotin hydrazide. (Lower) image of the Coomassie-stained gel.

Taken together, the PAGE and Western blot data suggest that aldehyde-tagged GFP expressed in the cytosol of HEK293 cells undergoes Cys to FGly conversion at low levels in the absence of exogenously expressed cytosolic FGE. We considered the possibility that aldehyde-tagged GFP was exposed to endogenous FGE during the process of cell lysis, despite our efforts to prevent disruption of the ER membrane by use of nondetergent buffers. To test this, we analyzed the isolated cytosolic fraction for the presence of BiP, a major ER-resident chaperone. Western blots confirmed that this protein was present at low levels in the cytosolic fraction, consistent with a small degree ER contamination (Fig. S4). Alternatively, there may exist a previously uncharacterized FGE activity in the cytosol of human cells. Likewise, FGE-like activities have been identified in E. coli and postulated in Caenorhabditis elegans, but the molecular identities of the corresponding enzymes have not yet been determined (26).

Conclusion

The aldehyde tag offers a practical and versatile method for site-specific chemical modification of secreted, membrane-associated, and cytosolic proteins expressed in mammalian systems. The reactions of aldehydes with aminooxy or hydrazide reagents are typically complete within 2 hours at 37 °C, and the resulting oximes and hydrazones, respectively, are quite stable under physiological condition. Moreover, many aminooxy- and hydrazide-functionalized reagents are commercially available; thus, the procedure requires only simple cloning steps to generate the necessary components for protein modification. With minimal optimization, we obtained recombinant proteins with >90% conversion of Cys to FGly. We expect that further manipulation of FGE expression levels and exploration of different cell lines will produce systems capable of even higher conversion. Importantly, we discovered that both 6-mer and 13-mer tags are viable for protein modification applications, although the 13-mer tag provides higher conversion efficiency of Cys to FGly in the IgG construct studied. It is notable that the FGE consensus sequence is highly restricted to Type I sulfatases, wherein the reactive aldehydes are buried in the active site. Analysis of the human genome sequence reveals very few proteins outside of this family with related sequence motifs (Table S2). Thus, even within unpurified cell lysates or conditioned media, recombinant aldehyde-tagged proteins are predicted to be the predominant species that label with aminooxy or hydazide reagents.

The presence of an unmodified Cys residue within the aldehyde tag sequence has some interesting consequences, both advantageous and potentially disadvantageous. In some cases, we found that the unmodified proteins formed disulfide-bound homodimers, which should enable their trivial separation from the corresponding monomeric FGly-containing proteins by size-exclusion methods. However, it is possible that unconverted Cys residues will interfere with folding of proteins possessing native disulfide bonds, although we observed no such detrimental effects in the cases of Fc and IgG.

The proteins reported in this work were labeled near their N or C termini. The natural substrates for FGE, Type I sulfatases, possess internal consensus sequences, suggesting that internal aldehyde tags may be recognized within heterologous proteins as well. An interesting extension of this work will be to insert aldehyde tag sequences into internal loops of recombinant proteins, perhaps at sites normally occupied by glycosylation or other posttranslational modifications. In this way, the aldehyde tag could be exploited to install a wide range of natural or unnatural posttranslational modifications anywhere on the surface of a folded protein. The precise chemical control offered by the aldehyde tag method should enable the development of new protein products for research and therapeutic purposes.

Materials and Methods

Protein Modification Reactions.

Chemical modification of aldehyde-tagged proteins was achieved by treating 4 μg of the target protein with 300 μM aminooxy- or hydrazide-functionalized probe [Alexa Fluor 488 C5-aminooxyacetamide (Invitrogen), biotin hydrazide (Sigma), or aminooxy-FLAG (16)] in labeling buffer [100 μM Mes (pH 5.5), 1% SDS] at room temperature for 2 h. After adding 4× SDS/PAGE loading buffer to each sample, reaction mixtures were resolved on Bis-Tris Criterion gels (4–12% or 12%; Bio-Rad). Protein loading was determined by Coomassie staining. Western blots were probed with either an α-biotin-HRP antibody (1:100,000 dilution; Jackson Immuno Research Laboratories) or with the α-FLAG M2-HRP antibody (1:3,000 dilution; Sigma) and developed by using SuperSignal West Pico Chemiluminescent Substrate. Aminooxy-FLAG (H₂NO-CH₂-CO-NH–(DYKDDDDK)–CO₂H) was synthesized via standard Fmoc-based solid-phase peptide synthesis protocols as previously described (27). The final residue added at the N terminus was (t-Boc-aminooxy)acetic acid followed by cleavage under standard conditions. The peptide was subsequently purified by C₁₈ reversed-phase HPLC.

Tryptic Digestion of Aldehyde-Tagged Fc.

After affinity purification using protein A/G agarose (protein A/G is an engineered construct possessing Fc binding domains from both protein A and protein G), each aldehyde-tagged Fc fragment was subjected to tryptic digestion. Each protein was first treated with Tris(2-carboxyethyl) phosphine hydrochloride (TCEP, 5 mM) at 37 °C for 30 min, followed by alkylation using iodoacetamide (20 mM) at room temperature for 30 min in the dark. The protein was boiled for 10 min, and trypsin (Promega) is added at 1:20 (mass ratio, enzyme/substrate) for overnight digestion. The completeness of protein digestion was verified by silver stain.

Quantification of the Conversion Efficiency of Cysteine to Formylglycine in the Aldehyde-Tagged Fc.

Peptide standards were custom synthesized by Biomer Technology. Heavy isotope-labeled peptides were used as the internal standards for quantification. The sequences of peptide standards are based on the predicted tryptic digestion products of Ald₆ and Ald₁₃. Specially, the sequences of the peptide standards for quantifying the N- and C-tagged Fc are ISLCTPSR and SLGTLCTPSR, respectively. The Leu residue of ISLCTPSR was replaced by an isotope-labeled counterpart, which increased the mass unit by 7 Da; whereas both Leu and Gly residues in SLGTLCTPSR were isotope labeled, which increased the peptide mass by 10 Da. The Cys of the peptide standards was carboxymethylated to have the same modification as the Fc digests.

LC-ESI-MS (Waters Micromass Q-TOF) and MALDI-TOF/TOF (ABI 4800) were used for quantifying the occupancy of the aldehyde tag on Fc. Individual Fc digest from 160 fmol of protein was mixed with 50 fmol of its corresponding internal standard, and the mixture was injected into LC-ESI-MS. MS/MS mode was applied in some cases to acquire the peptide sequence data. For MALDI-TOF/TOF analysis, both Fc digests and internal standards were diluted to 2 pmol/μL and mixed equally. A portion of the mixture (0.25 μl) was loaded onto a MALDI sample plate, followed by loading of the matrix solution (α-cyano-4-hydroxycinnamic acid in 60% of CH₃CN, 0.1% TFA). The signal of both Fc peptides and internal standards were within the linear range of detection by a given mass spectrometer. The peak ratio of a pair of unlabeled peptide and its labeled counterpart (of known quantity) shown in the mass spectrum allowed us to calculate the absolute amount of the unconverted peptide (unlabeled) derived from Fc. Finally the occupancy of the aldehyde tag on a given Fc was deduced from the quantity of the unconverted peptide and the total amount of Fc protein. An equation is shown below to illustrate the calculation. Occupancy of the aldehyde tag on Fc = 1 − [(the quantity of doped peptide standard × the isotope ratio*)/the quantity of Fc protein injected]

Labeling of Ald₁₃-TM Expressed on CHO Cells and Flow Cytometry Analysis.

CHO cells were transiently transfected with plasmids encoding Ald₁₃-TM or unmodified PDGFR-TM and hFGE (2:1 wt/wt). After 40 h, the cells were washed, collected, and resuspended in PBS containing 0.1% FBS (vol/vol) (pH 6.4) and 200-μL aliquots containing 5 × 10⁵ cells were distributed into a V-bottom 96-well plate. Biotin hydrazide was added at 1 mM final concentration, and the cells were incubated for 1 h at room temperature. Subsequently, the cells were washed 3 times with PBS containing 1% FBS (vol/vol) (PBS/FBS) and resuspended in 200 μl of PBS/FBS. Alex Fluor 488-streptavidin (Invitrogen) was added to the cells at final 2–3 μg/mL concentration. After incubating on ice for 25 min, the cells were washed 3 times with 200 μl of PBS/FBS and resuspended in 400 μl of PBS/FBS. Flow cytometry was performed by using a FACScalibur instrument (Becton Dickinson).

Analysis of Ald₆-Tagged Anti-Nogo R2 IgG Binding Activity.

HEK293F cells were transiently transfected with the human NgR2 in pCEP4 plasmid (Invitrogen) (22). After 48 h, the cells were collected by centrifugation and incubated in PBS containing 10% FBS (vol/vol) for 1 h. Cells were then centrifuged, resuspended in PBS/FBS, and 50-μL aliquots containing 5 × 10⁵ cells were distributed into a V-bottom 96-well plate. Alexa Fluor 488-conjugated Ald₆-IgG (anti-Nogo R2) was added at a final concentration of 6 μg/mL, and the cells were incubated for 1 h. Subsequently, the cells were washed twice with PBS containing 1% FBS (vol/vol) and resuspended in 400 μL of the same buffer. All steps were carried out on ice. Flow cytometry was performed by using a FACScalibur instrument (Becton Dickinson).

Fluorescent Microscopy Analysis of Labeled Membrane-Associated Proteins.

CHO cells were seeded at a density of 50,000 cells per milliliter into an 8-well microscopy tray and allowed to grow for 24 h. The cells were then transiently transfected with plasmids encoding Ald₁₃-TM or Ald₁₃-CD4, or with control plasmids, unmodified PDGFR-TM or Ald₁₃-CD4(C18A), along with hFGE (2:1 wt/wt). The cells were incubated for 48 h at 37 °C in a humidified 5% CO₂ atmosphere in Ham's F-12 medium supplemented with 10% FBS. The cells were washed (2 times in 300 μL of PBS) and incubated with aminooxy-Alexa Fluor 647 (50 μM) in PBS/FBS (pH 6.4) for 1 h. Hoechst 33342 dye (1 μg/mL, Invitrogen) was added, and after 1 min the cells were rinsed with PBS/FBS (pH 7.0, 3 times in 400 μL) and resuspended in 400 μL of the same buffer. The cells were analyzed by fluorescence microscopy at 25 °C using a Zeiss Axiovert 200M inverted microscope (Carl Zeiss MicoImaging Inc.) equipped with a 63 × 1.4 N.A. PlanApochromat oil-immersion lens. Image stacks containing 40 sections, spaced 0.1 μm apart, were acquired by using a CoolSNAP HQ CCD camera (Photometrics). The image stacks were digitally deconvolved by using the nearest-neighbor algorithm of Slidebook (Intelligent Imaging Innovations, Inc.).