Rapid and Modular Assembly of Click Substrates To Assay Enzyme Activity in the Newborn Screening of Lysosomal Storage Disorders.

Synthetic substrates play a pivotal role in the development of enzyme assays for medical diagnostics. However, the preparation of these chemical tools often requires multistep synthetic procedures complicating structural optimization and limiting versatility. In particular, substrates for enzyme assays based on tandem mass spectrometry need to be designed and optimized to fulfill the requirements to finally enable the development of robust diagnostic assays. In addition, isotope-labeled standards need to be prepared to facilitate accurate quantification of enzyme assay products. Here we report the development of a building block strategy for rapid and modular assembly of enzyme substrates using click chemistry as a key step. These click substrates are made up of a sugar moiety as enzyme responsive unit, a linker that can easily be isotope-labeled for the synthesis of internal standards, and a modifier compound that can readily be exchanged for structural optimization and analytical/diagnostic tuning. Moreover, the building block assembly eliminates the need for extensive optimization of different glycosylation reactions as it enables the divergent synthesis of substrates using a clickable enzyme responsive unit. The outlined strategy has been applied to obtain a series of synthetic α-l-iduronates and sulfated β-d-galactosides as substrates for assaying α-l-iduronidase and N-acetylgalactosamine-6-sulfate sulfatase, enzymes related to the lysosomal storage disorders mucopolysaccharidosis type I and type IVa, respectively. Selected click substrates were finally shown to be suitable to assay enzyme activities in dried blood spot samples from affected patients and random newborns.

Introduction

Lysosomal storage disorders (LSDs) are severe diseases, caused by genetic defects leading to a deficiency of lysosomal enzyme activity. The resulting progressive accumulation of macromolecular substrates is specific to each disorder and causes a gradual deterioration of cellular and tissue function.[1−3] Individual LSDs are considered as rare diseases with a round combined prevalence estimated at 1 per 7700 live births.[4] In contrast to general treatments for the symptoms of LSDs various therapeutic options have been developed dealing with the cause of the disease: (i) enzyme replacement therapy, (ii) chaperone-mediated enzyme enhancement, (iii) substrate reduction therapy, (iv) stem cell transplantation, and (v) hematopoietic stem cell gene therapy.[5−9] The development of those treatment options has led to an intense interest into newborn screening (NBS) of LSDs as early initiation of treatment often leads to a better outcome.[10] Different methods for NBS have been developed including (i) direct analysis of enzymatic activity, (ii) gene sequencing, (iii) direct measurement of enzyme abundance, and (iv) biomarker quantification.[10] While measurements of enzyme abundance and biomarker analysis are still at an early stage of development, and gene sequencing is limited by the lack of a complete list of pathogenic mutations, direct enzyme activity analysis has been extensively developed in recent years.[10] In their pioneering studies, Chamoles and co-workers assayed enzyme activities in rehydrated dried blood spots (DBS) using fluorometric and radiometric assays.[10−13] DBS analysis has several advantages over other methods as it requires only a few drops of blood, and samples can easily be sent in plastic envelopes at room temperature to other cities, countries, or specialized reference laboratories.[14]

Tandem mass spectrometry (MS/MS) has been developed as a diagnostic platform for early detection and screening of genetic disorders and many countries have implemented newborn screening using MS/MS.[15] Pioneered by Gelb and co-workers, MS/MS was applied for the newborn screening of several LSDs.[16−24] Most importantly, MS/MS changed the paradigm of analyzing one analyte per disorder as it allows the analysis of multiple enzymes in a single DBS punch (multiplexing). Several multiplex assays have been developed in recent years.[17,19,22] To simplify sample cleanup before flow injection into the MS/MS instrument several groups interfaced ultra-high-pressure liquid chromatography (UHPLC) with MS/MS. Two-dimensional chromatography (applying perfusion columns or turboflow online sample cleanup) was used to avoid the need for offline liquid–liquid and solid-phase extraction.[25−31]

In addition, fluorogenic substrates have been developed and used to assay several lysosomal enzymes.[11−14,32−41] Fluorometric methods that make use of digital microfluidics represent an emerging technology that is considered to be at least as effective as MS/MS for high-throughput screening of multiple LSDs.[42−44]

A key drawback in the development of enzyme assays (in particular MS/MS-based ones) is that the preparation of substrates often requires more than 10 synthetic steps.[10] Furthermore, as required for MS/MS methods, corresponding internal standards (in most cases isotope-labeled reference compounds of the assay products) need to be prepared. The labor- and cost-intensive development of substrates and internal standards impedes fast structural optimization (considering substrate stability, enzyme kinetics, MS ionization, in-source fragmentation, etc.)[26] with the aim to develop a robust assay. For instance, the development of improved substrates for the newborn screening of mucopolysaccharidosis (MPS) types I, II, and VI has been reported in 2014,[17] thus almost a decade after the first described synthetic substrate for MPS I,[23] and 6 years after the first structural optimization (13 synthetic steps and additional 2 steps to obtain substrate and internal standard, respectively).[16]

Deficiency of glycosidases is related to several LSDs and thus receiving increasing interest.[1,2,45] The preparation of substrates for the development of glycosidase assays (Figure a) involves glycosylation as a key step, a reaction that is known to be notoriously difficult and thus requiring extensive optimization.[46] Hence, this further complicates the screening and optimization of substrate structures when using a linear (or convergent) synthetic sequence.

Figure 1

(a) General concept of LC-MS/MS-based glycosidase assays. Glycosylated substrates are cleaved by the enzyme yielding the product (P) as analyte for subsequent analysis. An isotope-labeled internal standard (IS; here shown as deuterated product, P*) is required and used for quantification. (b) Divergent building block assembly of substrates using a single optimized glycosylation reaction and click chemistry to access a library of click substrates (CS). (c) Corresponding internal standards (IS) can easily be prepared using a single deuterated linker in combination with other building blocks.

In this light, we herein report the design of a strategy for the building block assembly of substrates using a click reaction as key step. A click tag is attached to a linker that can subsequently be conjugated to a library of compounds (modifiers) to tune the MS/MS properties of the final substrate. The sugar moiety is connected to a click aglycone using a single optimized glycosylation reaction, and the resulting building block is reacted with the clickable linker–modifier conjugates. This approach enables the divergent synthesis of click substrates (Figure b). Furthermore, a single deuterated linker can be used for facile preparation of the corresponding isotope-labeled internal standards (Figure c). While the click aglycone can be exchanged to optimize enzyme kinetics, variation of the sugar moiety leads to substrates for different enzyme assays using the same linker–modifier building blocks, which we finally show by the development of click substrates to assay a clinically relevant sulfatase. The versatility of this approach thus enables the synthesis, screening, and optimization of numerous substrates in short time. Individual building blocks can easily be exchanged without having to start a multistep synthesis from the beginning.

Results and Discussion

To develop and test the outlined strategy we started by focusing on substrates to assay α-L-iduronidase (IDUA). Deficiency of this lysosomal enzyme causes mucopolysaccharidosis type I (MPS I). As discussed above several substrates to assay MPS I have been developed including fluorogenic compounds and substrates optimized for MS/MS.[11,13,16,17,23] It is noteworthy that we did not aim for structural optimization of MPS I substrates to improve currently used assays that have been gradually optimized in the past decade, but to evaluate the general applicability of our click strategy for the divergent synthesis of enzyme substrates. IDUA is a lysosomal hydrolase responsible for the degradation of the glycosaminoglycans heparan sulfate and dermatan sulfate by cleaving an iduronic acid unit.[47] Hence, synthetic α-L-iduronidates are used to assay IDUA (Figure ). Chemical glycosylation to afford α-l-iduronidates represents a challenging, not well-studied reaction, which makes these compounds an ideal target for the outlined click strategy.

Figure 2

Cleavage of α-l-iduronates by IDUA.

For the synthesis of a readily accessible linker methyl 4-formylbenzoate (1) was reduced with NaBH4 to afford 2a, which was converted to 3a in an Appel reaction. Reduction with LiAlH4 yielding 4a and subsequent introduction of an azide moiety via nucleophilic substitution afforded clickable linker 5a(Figure a). This synthesis could easily be modified by using deuteride reagents to obtain the isotope-labeled linker 5b. A library of modifier compounds was prepared and used in this study for the synthesis of selected click markers (CM1–CM5) as shown in Figure a. Propargyl alcohol was used as a simple click aglycone and glycosylated by reaction with the known fluoroiduronyl donor 6(17) to afford the clickable iduronate 7. Copper-catalyzed azide alkyne cycloaddition (click chemistry)[48−51] was used for the divergent synthesis of click substrates. Selected compounds CS1–CS5 have been prepared via 8–12 (Figure b) by click reaction of 7 with CM1a–CM5a. The corresponding internal standards IS1–IS5 were prepared by reacting isotope-labeled click markers CM1b–CM5b with propargyl alcohol to obtain different analytical sets (CS, IS) for further investigations and the development of LC-MS/MS-based enzyme assays. The same synthetic procedures were used to obtain the products (P1–P5) of the enzymatic reactions (click of nonlabeled markers CM1a–CM5a with propargyl alcohol). These compounds were used for the development and tuning of analytical methods (synthesis not shown; for detailed description see the Supporting Information).

Figure 3

(a) Synthesis of nonlabeled and isotope-labeled azide-modified linkers 5a and 5b, respectively, and subsequent conjugation to modifier compounds (M) to obtain a library of click markers (CM). (b) Glycosylation of propargyl alcohol using iduronyl donor 6 and click assembly with CM1–CM5 to afford click substrates CS1–CS5. (c) Analytical sets of click substrates (CS) and corresponding internal standards (IS) for the development of LC-MS/MS-based α-l-iduronidase assays.

We have first used the click approach for the preparation of various substrates and internal standards applying chemically different modifier compounds. Selected compounds, including t-butylcarbonates (set 1), diethylcarbamates (set 2), and t-butyloxycarbonyl-protected (boc-protected) bis-carbamates (set 3), are shown in Figure c. These first generation sets were tested using recombinant human IDUA. Briefly, the substrates were reacted at a concentration of 1 mg/mL with increasing concentrations of pure IDUA (0, 25, 50, and 100 ng/mL) in IDUA assay buffer at pH 4.04 at 22 °C for 20 min. The isotope-labeled internal standards (IS) were added to the assay cocktail and used to quantify the product of the enzymatic cleavage of the click substrate. Enzymatic reactions were quenched by the addition of acetonitrile. Samples were centrifuged, and supernatants were diluted and analyzed by UHPLC-MS/MS, wherein mass spectrometric detection was performed using a triple quadrupole system operated in positive electrospray ionization (ESI) mode (for details see the Supporting Information). In these assays, we could show that the enzyme degrades all three substrates (CS1–CS3) as indicated by a linear correlation of IDUA enzyme activity and concentration (Figure a).

Figure 4

(a) Enzyme assays using recombinant human α-l-iduronidase (IDUA) and sets 1–5 of click substrate (CS1–CS5) and corresponding internal standard (IS1–IS5) (calculated specific activities are shown in μmol per min and mg enzyme). (b) Analysis of dried blood spots (DBS) using CDC control cards (QCL, QCM, QCH = quality control low, medium, high). (c) Analysis of DBS of confirmed MPS I patients (n = 9, anonymized) and random newborns (n = 88, anonymized). [****p < 0.0001.]

Subsequently, analyses of dried blood spots (DBS) were performed using quality control cards provided by the CDC (Center for Disease Control and Prevention, USA). Briefly, DBS cards were punched (3.2 mm diameter), and each spot was extracted with PBS buffer. The extracts were incubated with the click substrates in assay buffer (pH 4.04) for 22 h at 37 °C followed by quenching with acetonitrile and centrifugation after complete precipitation. The supernatants were diluted and analyzed by UHPLC-MS/MS using the same method as for IDUA assays as described above (for details see the Supporting Information). Data for selected CS/IS sets is shown in Figure b. In contrast to assays using recombinant human IDUA, sets 1 and 2 gave only poor results in the DBS assay, while set 3 was shown to be applicable to discriminate between low (QCL), medium (QCM), and high (QCH) enzyme concentration in DBS (Figure b). In general, DBS analyses differ from enzyme assays in terms of complexity and, most importantly, incubation time due to the much lower enzyme concentration in DBS. While IDUA assays were used to test and show enzyme activity, DBS analyses were performed to test our substrates in a clinically more relevant setup. The poor performance of sets 1 and 2 clearly indicates the need for a robust, rapid, and modular method for substrate synthesis as it is difficult to predict the performance in DBS assays and LC-MS/MS in general (considering solubility, stability, ionization efficiency, chromatographic performance, etc.). Based on these results sets 4 and 5 have been prepared that differ from set 3 only in the carbon-chain length of the modifier (Figure c, second generation sets). UHPLC-MS/MS assays have been performed using IDUA (Figure a) and DBS controls (Figure b) affording results similar to those of set 3.

Analysis of enzyme activity in DBS obtained from randomly selected newborns and 9 confirmed MPS I patients using CS/IS sets 3, 4, and 5 shows that the activity of IDUA in the patient samples is below the activity observed for random newborns (Figure c). Similar results were obtained for all three sets (except a higher variation in IDUA activity when using set 5). Even though these substrates do not achieve activities as described for recently developed optimized substrates for MPS I screening,[17] the building block approach can be used for further optimization, e.g., by exchanging the click aglycone to improve the kinetics of the enzymatic cleavage reaction.

In general, the major advantage of the click approach is that many substrates can be prepared by changing a single building block without the need for new synthetic strategies or optimized protocols. In addition, the corresponding internal standards can easily be prepared by using a single isotope-labeled linker. Hence, a group of substrates and IS that differ only in the carbon-chain length of the modifier can be prepared by exchanging the modifier building block as shown for sets 3–5. To investigate if such a group of similar substrates can be applied for the simultaneous analysis of different DBS (different patients) using a single UHPLC run, we performed a triplex assay, wherein “triplex” stands for 3 different samples screened for the activity of the same enzyme. DBS analyses were carried out as described above using the CS/IS sets 3, 4, and 5. Subsequently, the three separately incubated samples were combined and analyzed using a single injection (Figure a). All substrates and the corresponding internal standards could be separated in a 5 min UHPLC run (Figure b). To enable direct and better comparison of the data we have analyzed each DBS using the three different CS/IS sets and combined these samples before LC-MS/MS analysis. As shown in Figure c, we have been able to achieve results almost equivalent to those of the singleplex assays (Figure c), even though a higher variation from single- to triplex was observed for set 5 (Figure d). These results clearly indicate the feasibility of the approach that can potentially be extended by using additional substrates.

Figure 5

(a) Simultaneous analysis of three different samples using a single UHPLC-MS/MS run. (b) Chromatographic separation of CS/IS sets 3, 4, and 5 (including the products P3–P5 of the corresponding enzyme assays) using a 5 min gradient. (c) Triplex assay of three combined DBS samples using CS/IS sets 3, 4, and 5 (affected, n = 8; random, n = 23; ****p < 0.0001). (d) Analyzed DBS samples (random) in triplex vs singleplex assays (n = 23). [S = substrate, P* = IS, P = product.]

In addition to structural tuning and divergent synthesis of substrate libraries, a variety of substrates to assay different enzymes can be prepared by exchanging the clickable sugar moiety (enzyme responsive unit) and using the same synthetic protocols for rapid click assembly with already available linker–modifier building blocks. To demonstrate this key advantage, we focused on the development of substrates to assay N-acetylgalactosamine-6-sulfate sulfatase (GALNS). Deficiency of GALNS leads to accumulation of chondroitin-6-sulfate and keratan sulfate causing the lysosomal storage disorder mucopolysaccharidosis type IVa (Morquio syndrome, Figure a).[52,53] GALNS has been assayed in dried blood spots to screen for MPS IVa using fluorometric and LC-MS/MS-based methods.[20,26,54−56]

Figure 6

(a) Deficiency of GALNS leads to accumulation of chondroitin-6-sulfate and keratan sulfate causing MPS IVa. (b) Click assembly of GALNS substrates and internal standards using clickable galactose derivatives and already available linker–modifier building blocks. (c) Synthesis of GALNS sets 6–8 of click substrates (CS6–CS8) and corresponding internal standards (IS6–IS8) using click markers CM3a–CM5a and isotope-labeled CM3b–CM5b, respectively. (d) Analysis of dried blood spots (DBS) using CDC control cards. (e) Analysis of DBS of confirmed affected patients (n = 9, anonymized) and random newborns (n = 116, anonymized). [****p < 0.0001.]

Similar to the overall strategy as outlined in Figure , we aimed to use clickable galactose building blocks to prepare sulfated click substrates and the respective nonsulfated internal standards by using already available click marker (CM) building blocks (linker, deuterated linker, modifier; Figure b). Starting from 1-O-propargyl-β-d-galactose (13, click galactose) we have prepared the sulfated galactose building block 14 by protecting group manipulations (TIPS-protection, acetylation, removal of TIPS) and chemical sulfation on C-6. Click assembly of the alkyne-modified galactose moieties 14 and 13 with click markers CM3a–CM5a and CM3b–CM5b, respectively, afforded click substrates CS6–CS8, internal standards IS6–IS8 (Figure c), and the non-isotope-labeled products P6–P8 as reference compounds for tuning the MS instrument (see the Supporting Information).

These GALNS substrates were tested in DBS analyses using CDC control cards (similar to assaying IDUA in DBS as described above) showing that all three sets can be used to discriminate between low (QCL), medium (QCM), and high (QCH) enzyme concentration (Figure d). Moreover, sets 6–8 gave highly similar results showing independence of the assay from the length of the modifier. As the observed activities of >10 μM/h are significantly higher than or at least similar to previously reported data using MS/MS-based[20,26] or fluorometric methods,[54,55] we aimed to test our GALNS click substrates in a clinically more relevant setup. Therefore, we have analyzed DBS of 9 confirmed MPS IVa patients in comparison to randomly selected newborns (all anonymized). DBS cards were punched, and each spot was incubated in assay buffer (pH 4.04) containing substrate and internal standard for 22 h at 37 °C followed by quenching with acetonitrile and centrifugation. The supernatants were diluted and analyzed by UHPLC-MS/MS (see the Supporting Information). We obtained mean activities of 0.76 μM/h (affected patients) and 13.5 μM/h (random newborns), and thus an activity ratio (normal/affected) of >17 with a mean absolute difference of 12.7 μM/h between normal and affected newborns.

Not only could we thus show the modularity of the click approach by designing substrates for a different enzyme by changing a single building block, but we also expanded our strategy to the assaying of sulfatases by using a clickable sulfated sugar moiety.

Conclusions

The presented approach for building block assembly of click substrates (CS) and corresponding internal standards (IS) was shown to be suitable for the divergent, versatile, and modular synthesis of chemical tools and the development of diagnostic LC-MS/MS-based enzyme assays. A library of substrates to screen for MPS I was prepared and investigated in enzyme assays and DBS analyses. Selected substrates have been shown to be suitable to determine the enzymatic activity in patient samples. By changing a single building block, we were able to show (i) rapid synthesis of a group of similar CS/IS sets for the simultaneous analysis of DBS using a single chromatographic run, and (ii) the development of MPS IVa substrates using a sulfated clickable galactose derivative and already available linker–modifier building blocks. These GALNS click substrates were evaluated using quality control cards and shown to be suitable to assay enzyme activities in dried blood spot samples from affected patients and random newborns.

Overall, we were able to demonstrate the advantages and further potential of the click approach and are thus convinced that this strategy will accelerate and contribute to the development of diagnostic tools, LC-MS/MS-based enzyme assays, and screening methods.