+ All Categories
Home > Documents > Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection...

Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection...

Date post: 19-Oct-2019
Category:
Upload: others
View: 12 times
Download: 1 times
Share this document with a friend
10
Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li 1 , Juntao Yu 2 , Jingyi Zhao 1 , Jie Wang 2 , Siqi Zheng 1 , Shixian Lin 1 , Long Chen 1 , Maiyun Yang 1 , Shang Jia 1 , Xiaoyu Zhang 3 and Peng R. Chen 1,2 * Employing small molecules or chemical reagents to modulate the function of an intracellular protein, particularly in a gain- of-function fashion, remains a challenge. In contrast to inhibitor-based loss-of-function approaches, methods based on a gain of function enable specific signalling pathways to be activated inside a cell. Here we report a chemical rescue strategy that uses a palladium-mediated deprotection reaction to activate a protein within living cells. We identify biocompatible and efficient palladium catalysts that cleave the propargyl carbamate group of a protected lysine analogue to generate a free lysine. The lysine analogue can be genetically and site-specifically incorporated into a protein, which enables control over the reaction site. This deprotection strategy is shown to work with a range of different cell lines and proteins. We further applied this biocompatible protection group/catalyst pair for caging and subsequent release of a crucial lysine residue in a bacterial Type III effector protein within host cells, which reveals details of its virulence mechanism. T he discovery and development of small molecules and other non-invasive approaches to activate a protein of specific inter- est in living cells has drawn considerable attention in recent years 1–6 . Albeit exceedingly difficult, the identification of small-mol- ecule protein activators remains a central task because of their remarkable advantages in facilitating a gain-of-function study rather than the widely adopted, inhibitor-based loss-of-function study of target proteins 1 . For example, a panel of small-molecule activators, the majority of which are allosteric activators 7 , has been developed for switching on the intrinsic activity of enzymes and provides a way to ascertain the sufficiency of an enzyme in trig- gering a specific signalling pathway or cellular response. Another intriguing example is the chemical rescue strategy in which an exter- nal small molecule is utilized to restore the wild-type activity of a protein that contains active-site mutations, which enables the dis- section of its specific roles in modulating intracellular signal trans- duction 8,9 . However, typically these small-molecule protein activators are identified by high-throughput screening or discovered by serendipity, and are thus not directly transferrable to other pro- teins 1 . Alternatively, mechanism-based photocaged versions of proteins or small effector molecules have been developed, and allow spatial and temporal activation of a protein under living con- ditions 3,10,11 . Such methodologies rely on light-mediated removal of the caging group as a way to modulate the functionality or localiz- ation of the target proteins in vitro and in vivo. Among these approaches, direct caging of proteins via a genetically encoded unnatural amino acid (UAA) bearing a photocleavable group allows control of protein activity in a site-specific fashion 12–14 . However, ultraviolet irradiation on live cells may trigger surface- receptor internalization 15 , alter the intracellular signalling networks and cause additional cytotoxic effects (Supplementary Fig. 1). In addition, the poor penetration capability renders these photocaged UAAs incapable of being further developed for utilization in deep tissues or intact animals. Inspired by the aforementioned chemical rescue approach, we envisioned that direct blockage of an essential residue on a protein of interest by a cleavable group may generate a ‘chemically caged’ protein with the wild-type activity temporally switched off. The subsequent addition of a membrane-permeable and biocompatible cleavage reagent may drive the elimination of this caging group and thus restore the native functionality to a protein within its intracellular context. Such a ‘chemical decaging’ strategy may offer an attractive technique based on small molecules and generally applicable for manipulating the activity of proteins within living systems (Fig. 1a). Lysine is a key amino acid in protein posttranslational modifi- cations (PTMs) and transduction of intracellular signals 16 , and it also plays essential roles in various enzymes such as protein kinases 17 . We are interested in developing a biocompatible chemical control strategy to cage and release the 1-amine from a lysine, which may allow modulation of the function, structure and/or localization of a protein that relies on a key lysine residue. In our efforts towards searching for a desired biocompatible cleavage group to modulate protein activity chemically via the liberation of free lysine, we turned our attention to allyl and propargyl, two commonly used protecting groups for alcohol/amine and carbonate/carbamate, that can be deprotected by organometallic catalysts under mild or living conditions 18–24 (Fig. 1b). Herein we report the development of a palladium-mediated chemical decaging strategy to control lysine-dependent activation of intracellular proteins. By using bio- compatible palladium catalysts identified from this study, we suc- cessfully liberated the 1-amine from a propargyloxycarbonyl (Proc)-‘caged’ lysine analogue (3, Fig. 1c) that was genetically and site-specifically incorporated into an intracellular protein, allowing facile restoration of the wild-type activity to the chemically caged protein with its biological roles investigated within living cells. Results and discussion Identifying palladium catalysts for the biocompatible and efficient deprotection of caged lysine. We started by synthesizing two fluorogenic rhodamine derivatives, Proc-protected rhodamine 1 Synthetic and Functional Biomolecules Center, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Bioorganic Chemistry and Molecular Engineering of Ministry of Education, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China, 2 Peking-Tsinghua Center for Life Sciences, Beijing 100871, China, 3 College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000, China. *e-mail: [email protected] ARTICLES PUBLISHED ONLINE: 16 MARCH 2014 | DOI: 10.1038/NCHEM.1887 NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry 352 © 2014 Macmillan Publishers Limited. All rights reserved.
Transcript
Page 1: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

Palladium-triggered deprotection chemistry forprotein activation in living cellsJie Li1, Juntao Yu2, Jingyi Zhao1, Jie Wang2, Siqi Zheng1, Shixian Lin1, Long Chen1, Maiyun Yang1,

Shang Jia1, Xiaoyu Zhang3 and Peng R. Chen1,2*

Employing small molecules or chemical reagents to modulate the function of an intracellular protein, particularly in a gain-of-function fashion, remains a challenge. In contrast to inhibitor-based loss-of-function approaches, methods based on again of function enable specific signalling pathways to be activated inside a cell. Here we report a chemical rescue strategythat uses a palladium-mediated deprotection reaction to activate a protein within living cells. We identify biocompatibleand efficient palladium catalysts that cleave the propargyl carbamate group of a protected lysine analogue to generate afree lysine. The lysine analogue can be genetically and site-specifically incorporated into a protein, which enables controlover the reaction site. This deprotection strategy is shown to work with a range of different cell lines and proteins. Wefurther applied this biocompatible protection group/catalyst pair for caging and subsequent release of a crucial lysineresidue in a bacterial Type III effector protein within host cells, which reveals details of its virulence mechanism.

The discovery and development of small molecules and othernon-invasive approaches to activate a protein of specific inter-est in living cells has drawn considerable attention in recent

years1–6. Albeit exceedingly difficult, the identification of small-mol-ecule protein activators remains a central task because of theirremarkable advantages in facilitating a gain-of-function studyrather than the widely adopted, inhibitor-based loss-of-functionstudy of target proteins1. For example, a panel of small-moleculeactivators, the majority of which are allosteric activators7, hasbeen developed for switching on the intrinsic activity of enzymesand provides a way to ascertain the sufficiency of an enzyme in trig-gering a specific signalling pathway or cellular response. Anotherintriguing example is the chemical rescue strategy in which an exter-nal small molecule is utilized to restore the wild-type activity of aprotein that contains active-site mutations, which enables the dis-section of its specific roles in modulating intracellular signal trans-duction8,9. However, typically these small-molecule proteinactivators are identified by high-throughput screening or discoveredby serendipity, and are thus not directly transferrable to other pro-teins1. Alternatively, mechanism-based photocaged versions ofproteins or small effector molecules have been developed, andallow spatial and temporal activation of a protein under living con-ditions3,10,11. Such methodologies rely on light-mediated removal ofthe caging group as a way to modulate the functionality or localiz-ation of the target proteins in vitro and in vivo. Among theseapproaches, direct caging of proteins via a genetically encodedunnatural amino acid (UAA) bearing a photocleavable groupallows control of protein activity in a site-specific fashion12–14.However, ultraviolet irradiation on live cells may trigger surface-receptor internalization15, alter the intracellular signalling networksand cause additional cytotoxic effects (Supplementary Fig. 1). Inaddition, the poor penetration capability renders these photocagedUAAs incapable of being further developed for utilization in deeptissues or intact animals. Inspired by the aforementioned chemicalrescue approach, we envisioned that direct blockage of an essential

residue on a protein of interest by a cleavable group may generate a‘chemically caged’ protein with the wild-type activity temporallyswitched off. The subsequent addition of a membrane-permeableand biocompatible cleavage reagent may drive the elimination ofthis caging group and thus restore the native functionality to aprotein within its intracellular context. Such a ‘chemical decaging’strategy may offer an attractive technique based on small moleculesand generally applicable for manipulating the activity of proteinswithin living systems (Fig. 1a).

Lysine is a key amino acid in protein posttranslational modifi-cations (PTMs) and transduction of intracellular signals16, and italso plays essential roles in various enzymes such as proteinkinases17. We are interested in developing a biocompatible chemicalcontrol strategy to cage and release the 1-amine from a lysine, whichmay allow modulation of the function, structure and/or localizationof a protein that relies on a key lysine residue. In our efforts towardssearching for a desired biocompatible cleavage group to modulateprotein activity chemically via the liberation of free lysine, weturned our attention to allyl and propargyl, two commonly usedprotecting groups for alcohol/amine and carbonate/carbamate,that can be deprotected by organometallic catalysts under mild orliving conditions18–24 (Fig. 1b). Herein we report the developmentof a palladium-mediated chemical decaging strategy to controllysine-dependent activation of intracellular proteins. By using bio-compatible palladium catalysts identified from this study, we suc-cessfully liberated the 1-amine from a propargyloxycarbonyl(Proc)-‘caged’ lysine analogue (3, Fig. 1c) that was genetically andsite-specifically incorporated into an intracellular protein, allowingfacile restoration of the wild-type activity to the chemically cagedprotein with its biological roles investigated within living cells.

Results and discussionIdentifying palladium catalysts for the biocompatible andefficient deprotection of caged lysine. We started by synthesizingtwo fluorogenic rhodamine derivatives, Proc-protected rhodamine

1Synthetic and Functional Biomolecules Center, Beijing National Laboratory for Molecular Sciences, Key Laboratory of Bioorganic Chemistry and MolecularEngineering of Ministry of Education, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China, 2Peking-Tsinghua Centerfor Life Sciences, Beijing 100871, China, 3College of Chemistry and Chemical Engineering, Lanzhou University, Lanzhou 730000, China.

*e-mail: [email protected]

ARTICLESPUBLISHED ONLINE: 16 MARCH 2014 | DOI: 10.1038/NCHEM.1887

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry352

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 2: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

110 (Proc-R110, 5) and allyloxycarbonyl (Aloc)-‘caged’ rhodamine110 (Aloc-R110, 6)18, as the reporters to assess the efficiency ofpalladium-mediated depropargylation and deallylation reactions,respectively (Fig. 2a). These two probes are virtually non-fluorescent until the conversion into a highly fluorescentcompound (7, F¼ 0.91, excitation (Ex)¼ 485 nm, emission(Em)¼ 520 nm) via the cleavage of protection groups (Fig. 2a).These fluorogenic substrates can link the deprotection yielddirectly with a fluorescent readout, which may offer a facileapproach for an assessment of the reaction efficiency catalysed bydifferent palladium species in vitro and in vivo. As bulky andnegatively charged ligands may have membrane-permeabilityissues, simple and single-component palladium sources arepreferred to facilitate the cellular entry, and may also be less toxicthan many palladium complexes that contain phosphineligands25,26. Therefore, as shown in Fig. 2b, a total of six simple

and air-stable palladium species were surveyed. Among these, C4is an air-stable palladium(0) precursor27,28 and C5 is a simple andstable Pd(II) compound, but is readily converted into Pd(0)species on nucleophilic (for example, water) attack throughreductive elimination27,29. In addition, three commonly used andwater-soluble ligands (L1, L2 and L3) were used in the complexwith simple palladium sources for comparison. The eliminationefficiency catalysed by each of the palladium species was evaluatedsystematically by analysing the enhanced fluorescence(Supplementary Method 2). Interestingly, all six simple palladiumcatalysts were found to catalyse the depropargylation reaction onProc-R110 more efficiently than the deallylation reaction on Aloc-R110 (Fig. 2c), which is in line with previous reports that strongreducing reagents and/or phosphine ligands are necessary topromote the simple deallylation reaction mediated by a palladiumsource19. We found that two simple palladium compounds,allyl2Pd2Cl2 (C5, allylpalladium(II) chloride dimer) and Pd(dba)2(C4) (dba¼ dibenzylidene acetone), were the most-efficientcatalysts for both types of elimination reactions (Fig. 2c,Supplementary Fig. 2). In contrast, the addition of water-solubleligand L1 to C2 increased the deallylation efficiency to a certainextent, but not the depropargylation efficiency. In addition,although ligands L2 and L3 have been used previously to facilitatethe palladium-catalysed Suzuki–Miyaura and copper-freeSonogashira cross-coupling reactions on proteins26,30, respectively,both of these ligands were found to decrease significantly thePd(OAc)2(C1)-mediated depropargylation on Proc-R110.

To confirm further that these identified palladium species candepropargylate the Proc-protected lysine amino acid effectively,and that this reaction is catalytic, we synthesized N1-propargyloxy-carbonyl-L-lysine (Proc-Lys) (3, Fig. 1c), as well as 9-fluorenyl-methoxycarbonyl (Fmoc)-protected Proc-Lys (Proc-(Fmoc)-Lys;structure shown in Supplementary Fig. 3). The reaction yields ofboth C4- and C5-mediated depropargylation on Proc-Lys andProc-(Fmoc)-Lys were analysed by liquid chromatography–massspectroscopy (LC-MS), which showed that a catalytic loading ofeither of these two palladium species in the absence of anyadditional reagents can release free lysine with high efficiency(Fig. 2d and Supplementary Table 1). In particular, a 10% loadingof C4 or C5 in PBS buffer (pH 7.4) at 37 8C drove the depropargy-lation reaction to .80% completion (Fig. 2d). In contrast, neither ofthese palladium catalysts can efficiently liberate the Aloc-protectedlysine analogue, N1-allyloxycarbonyl-L-lysine (Aloc-Lys, 4, Fig. 1c)or the Fmoc-protected Aloc-Lys (Aloc-(Fmoc)-Lys, structureshown in Supplementary Fig. 3) under the same conditionswithout additional reducing agents (Fig. 2d and SupplementaryTable 1). Furthermore, our additional experiments indicated thatthe two most-efficient palladium compounds, allyl2Pd2Cl2 andPd(dba)2, may catalyse the depropargylation reaction via Pd(0)species with free amine and hydroxyacetone as the final products31

(Supplementary Fig. 4). Taken together, our results indicate that theProc group can be deprotected more effectively than the Aloc groupby simple palladium catalysts. As the bioorthogonality of the alkynemoiety is well-established, this Proc group and our identified palla-dium reagents may serve as a biocompatible ‘protection group/cata-lyst pair’ for lysine activation under living conditions.

Protein-based verification of palladium reagents for deprotectionof lysine. Next, we aimed to evaluate systematically theaforementioned palladium reagents for propargyl carbamatecleavage in the context of a carrier protein. Proc-Lys is an analogueof pyrrolysine (Pyl, 2, Fig. 1c), the 22nd naturally occurring aminoacid encoded by a Pyl-tRNA synthetase (PylRS)/tRNACUA

Pyl pairin Archaea species32. This Pyl-based system has been adapted forincorporation of various lysine-derived UAAs in response to an in-frame amber codon on target proteins in various living

NH2H2N

COOHLys (1)

HN

OCOOH

NH2N

Pyl (2)

HN O

OCOOH

H2N

Proc-Lys (3)

HN O

O

H2N

COOH

Aloc-Lys (4)

c

bR Y

R Y

Pd0/PdII/IV

H2O

Pd0

Nuc–, H2ORYH

RYH

Propargyl group

Allyl group

Depropargylation

Deallylation

Y = NH, O

HN NH2

X

a

X

'Chemicaldecaging'

: Caging group

Inactive protein Active protein

Cell membrane

Small moleculesor catalysts

Figure 1 | A chemical decaging strategy for protein activation in living

cells. a, A key amino acid (for example, a lysine residue in this study) on a

protein of interest may be site-specifically replaced by its genetically

encoded analogue that bears a caging group ‘X’, rendering the protein

inactive. Intracellular addition of an external small molecule or biocompatible

catalyst would trigger a deprotection reaction to remove this caging group,

and so release the decaged amino-acid side chain with restored protein

activity inside living cells. b, Deprotection chemistry as the potential

chemical decaging method. In contrast to the deprotection of the propargyl

group, which can be carried out with palladium catalysts in all the typical

oxidation states (0, þ2 and þ4), the allyl group can only be deprotected by

Pd(0) species. c, Structures of lysine (Lys, 1), pyrrolysine (Pyl, 2), N1-

propargyloxycarbonyl-L-lysine (Proc-Lys, 3) and N1-allyloxycarbonyl-L-lysine

(Aloc-Lys, 4). The liberation of lysine’s 1-amine from 3 and 4 can be

controlled by palladium-mediated propargyl carbamate cleavage and allyl

carbamate cleavage reactions, respectively.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887 ARTICLES

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry 353

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 3: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

organisms13,33–35. Proc-Lys was first site-specifically incorporated intogreen fluorescent protein (GFP) at residue Asn149 to produce GFP-N149-ProcLys in Escherichia coli cells36 (Supplementary Method 5)and the fidelity of this incorporation was verified by massspectrometry (Supplementary Fig. 5).

Five of the relatively more-efficient palladium species in Fig. 2bwere surveyed for depropargylation of Proc-Lys from GFP-N149-ProcLys in mild conditions that resemble those met in livingsystems: PBS buffer, pH 7.4, room temperature, no N2 protection(Fig. 3a). LC-MS analysis was employed to compare the molecularweight (MW) of intact GFP-N149-ProcLys protein before and afterpropargyl carbamate cleavage (Supplementary Fig. 6). Within onehour, a MW loss of 82 Da, which corresponds to the removal ofthe Proc group from Proc-Lys, was clearly detectable fromprotein samples treated by all the palladium reagents (Fig. 3b,Supplementary Fig. 6). As the catalytic activity of each palladiumcompound is proportional to the relative abundance of theGFP-N149-Lys product in comparison with the original GFP-N149-ProcLys protein, the cleavage efficiency can be calculatedaccording to the equation shown in Supplementary Fig. 6.Consistent with the aforementioned results from fluorogenic repor-ters (Fig. 2c), C5 (allyl2Pd2Cl2) was found to exhibit the highestdeprotection efficiency followed by C4 (Pd(dba)2), with the depro-tection yields reaching 90 and 78%, respectively (Fig. 3a,Supplementary Fig. 6). Furthermore, the reaction yield and thespecific reaction site on GFP-N149-ProcLys were validated byLC-MS-MS analysis on digested peptides that contained theProc-Lys residue (Supplementary Fig. 7), which showed a depro-tection yield of 86% by C5. Additionally, the impact of the

water-soluble phosphine ligand L1 was investigated, whichdecreased the efficiency for both the C4- and C5-mediated elimin-ations (Fig. 3a). Also, unlike Cu(I) ions utilized in the CuAAC reac-tion, working concentrations of these palladium species did notgenerate detectable reactive oxidative species (ROS) in the presenceof reducing agents (Supplementary Fig. 8), and they also exhibitednegligible damaging effects on GFP as well as on firefly luciferase, amore weakly folded protein (Supplementary Figs 9 and 10). Takentogether, we demonstrated that our identified palladium catalysts,allyl2Pd2Cl2 and Pd(dba)2, are highly efficient propargyl carbamatecleavage reagents compatible with intact proteins.

Palladium-mediated propargyl carbamate cleavage in living cells.To examine the compatibility of our identified palladiumcompounds within living cells, we first characterized theircytotoxicity by MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assay.Allyl2Pd2Cl2 and Pd(dba)2 exhibited negligible toxicity to HeLacells (three hours incubation, Supplementary Fig. 11) at aconcentration of 10 mM. The nucleus and membrane integrity ofthe cells after palladium treatment were also confirmed byHoescht 33342 nuclear staining and propidium iodide staining,respectively (Supplementary Fig. 12). Next, to exclude thepossibility that the potential cell-damaging effects stimulated bypalladium compounds were not detectable until after a longerincubation time, we extended the toxicity analysis of these twoidentified palladium reagents to a longer timescale. First, HeLacells and another five different cell lines (HEK293T, CHO, Caco-2, A549 and NIH3T3) were treated with 10 mM allyl2Pd2Cl2 or

OHN

HN

O

O

O

O

O

O

OH2N NH2

+

O

O

OHN

HN

O

O

O

O

O

OP

NaO3S

SO3Na

SO3Na

NH2

ONaNaO

N NN N

N

ONaNaO

a b Pd species

Pd(OAc)2 Pd(NO3)2 Na2PdCl4

Pd(dba)2 Allyl2Pd2Cl2 K2PdCl6

Water-soluble ligands

C1

[Pd]Proc-R110 (5)

Aloc-R110 (6)

R110 (7)(Ex = 485 nm, Em = 520 nm)

C2 C3

C4 C5 C6

L1

L2 L3

*Yields were determined by LC analysis (Supplementary Methods 3).†PBS buffer (pH = 7.4), 37 °C, 8 h.

Substrate† ProductEntry Catalyst (10 mol%) Yield (%)*

1

2

3

4

Lys84

Proc-LysC4

C5 82

0

4,000

8,000

12,000

16,000

Aloc-R110Proc-R110

C1 C2 C3 C4 C5 C6

C2 + L1

C1 + L2

C1 + L3

Contro

l

Flu

ores

cenc

e in

tens

ity (

a.u.

)c d

Lys26

Aloc-LysC4

C5 22

Figure 2 | Catalyst screening for palladium-mediated deprotection reactions. a, Structures of Proc-caged and Aloc-caged fluorogenic rhodamine 110 dyes

(Proc-R110, 5; Aloc-R110, 6) that can be converted readily into the highly fluorescent R110 (7) (Ex¼ 485 nm, Em¼ 520 nm) on deprotection triggered by

palladium catalysts. The fluorescence turn-on property of these two fluorophores can be used to screen catalysts in depropargylation and deallylation

reactions, respectively. b, The chemical formula and structure of palladium species and the water-soluble ligands used in this study. The water-insoluble

palladium species were prepared as DMSO stock solutions and diluted into water before usage. c, Measurement of the enhanced fluorescence by different

palladium species using fluorogenic probes 5 (red) and 6 (blue) as the substrates. Error bars represent+1 standard deviations (s.d.) from three independent

experiments. d, Confirmation of the catalytic property of the identified highly efficient palladium catalysts in the deprotection of Proc-Lys. Deprotection of

Aloc-Lys by the same palladium catalysts was used for comparison.

ARTICLES NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry354

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 4: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

Pd(dba)2 for three hours and then cultured for another 24 hours infresh media before being analysed by MTS assay, and all showedhigh viability (.90%) (Fig. 4a). Additionally, the incubation timesof palladium compounds were extended to 24 hours and 48 hoursbefore being subjected to MTS assay to confirm thebiocompatibility on a long timescale37 (Supplementary Fig. 13).Furthermore, the intracellular uptake and distribution ofallyl2Pd2Cl2 was analysed by inductively coupled plasma massspectrometry (ICP-MS) after treating cells with 10 mM catalyst for180 minutes. Palladium compounds were found to enter HeLacells effectively and distribute through both cytoplasm andnucleus (Fig. 4b). Finally, we also confirmed that allyl2Pd2Cl2 didnot generate ROS inside HeLa cells (Supplementary Fig. 14).

To ascertain whether these palladium reagents were able to carryout effectively the cleavage of propargyl carbamate inside cells,Pd(dba)2 or allyl2Pd2Cl2 was added to live HeLa cells preincubatedwith fluorogenic compound 5, and each was found to convert 5effectively into the highly fluorescent compound 7 inside HeLacells (Supplementary Figs 15 and 16). Time-lapse fluorescenceimaging of this process further indicated that the reaction reachedsaturation within about 180–240 minutes after the addition of

palladium reagents (Supplementary Fig. 15a). To quantify theturn-on ratio and optimize the concentration of palladium reagentsfor in vivo depropargylation, live HeLa cells after depropargylationreactions were subjected to flow cytometric analysis or in-cell fluor-escence assay (Supplementary Method 10). Both results indicatethat 10 mM Pd loading was sufficient for in vivo depropargylationwith high activity and low toxicity (Supplementary Fig. 16). Inaddition, to demonstrate the generality of this depropargylationchemistry in different cell lines, palladium-mediated deprotectionof 5 was monitored in the aforementioned six cell lines by in-cell flu-orescence assay and showed that 5 was readily converted into fluor-escent compound 7 in all cases (Fig. 4c). Fluorescence imaging wasalso conducted to verify the intracellular generation of 7 as well asthe integrity of the nucleus and membrane in all these cell types(Supplementary Fig. 17). Taken together, the results for our twoidentified depropargylation palladium reagents establish satisfyingbiocompatibility and elimination efficiency in living cells.

Palladium-mediated lysine liberation on intracellular proteins.To apply our strategy of palladium-mediated lysine liberation toproteins within living cells, we first extended the Proc-Lys

78

Substrate ProductEntry Catalyst Yield (%)

1

2

3

4

5

6

C1

7

4

9

11

90

C4

76

57

C2

C3

C5

C5 + P

C4 + P

a

b

–Met27,606

General conditions: proteins (10 µM), catalysts (100 µM), PBS buffer(pH 7.4), 25 °C, air, 1 h. Yields were determined by TOF-MS analysis(Methods in Supplementary Information). P = P(PhSO3Na)3 (200 µM).

–Met27,688

27,737

27,000 28,000 29,000

100

0

Inte

nsity

(%

)

27,819

After decagingBefore decaging

Mass

GFP-N149-ProcLys

GFP-N149-Lys

Allyl2Pd2Cl2

700 750 800 850 900 950

020406080

100

694.

4584

712.

2236

730.

9424

750.

6727

771.

4896

793.

5066

816.

8118

841.

5405

867.

7945

895.

7517

925.

5789

700 750 800 850 900 950

0

20

4060

80

100

696.

4832

714.

3167

733.

0878

752.

8276

773.

7581

795.

8363

819.

2233

844.

0072

870.

3502

898.

3933

928.

3038

GFP-N149-ProcLys

m/z

m/z

Figure 3 | Identification of biocompatible and efficient palladium reagents for propargyl carbamate cleavage on intact proteins. a, Proc-Lys was site-

specifically incorporated at residue Asn 149 on GFP to generate GFP-N149-ProcLys, which was used as a model protein for screening palladium-based

propargyl carbamate cleavage reagents. All the reactions were allowed to proceed for one hour followed by yield determination from each palladium reagent

with LC-TOF-MS. Reaction conditions: PBS buffer (pH 7.4) at room temperature without N2 protection for all entries. TOF, time of flight. b, Mass spectrometry

characterization of the propargyl carbamate cleavage reaction on GFP-N149-ProcLys catalysed by C5 (allyl2Pd2Cl2) under the same conditions as entry 3 in a.

The deconvoluted spectra (left) with their original ESI-MS results (right) are shown. GFP-N149-ProcLys (black), expected mass 27,819 Da, found mass 27,819

Da (major) and 27,688 Da (minor). GFP-N149-Lys (the depropargylated product, red), expected mass 27,737 Da, found mass 27,737 Da (major) and 27,606

Da (minor).The minor peak corresponds to the same protein (the major peak) with the N-terminal Met (–132 Da) posttranslationally cleaved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887 ARTICLES

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry 355

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 5: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

incorporation method to proteins expressed in mammalian cells.Following a protocol shown in Methods, we expressed theresultant protein GFP-Y40-ProcLys carrying a Proc-Lys UAA atresidue Y40. HeLa cells that expressed this GFP variant were thentreated with allyl2Pd2Cl2 (10 mM) to initiate the depropargylationreaction. To measure the depropargylation efficiency onintracellular proteins, an in-gel fluorescence assay was developedby taking advantage of the terminal alkyne moiety on Proc-Lys asa ‘clickable’ handle (details in Methods). As shown in Fig. 4d,after cells that harboured GFP-Y40-ProcLys were treated withpalladium reagents, the unreacted portion of the protein can belabelled by an azide-containing fluorophore through CuAAC. Thesubsequent in-gel fluorescence analysis could be used to assess thedeprotection efficiency. Following this protocol, we quantified theallyl2Pd2Cl2-mediated depropargylation reaction on GFP-Y40-ProcLys (10 mM Pd, three hours) which showed an efficiency of31% (Fig. 4e,f ). The reliability of this method was validated by

LC-MS-MS analysis on digested peptides that contained the Proc-Lys residue (Fig. 4f, Supplementary Fig. 18). The measureddeprotection efficiency from this MS analysis reached 27%.Together, these two different approaches reveal a similar efficiencyfor palladium-mediated depropargylation reaction on a Proc-Lysresidue from intracellular proteins, and the slight variations mayresult from intrinsic differences in the techniques. This observeddecline in elimination efficiency for intracellular proteinscompared with purified proteins is probably because of (1) thelow concentration of the target protein, which decreases theoverall conversion, and (2) the nonspecific absorption ofpalladium species by other biomolecules that largely decrease theeffective concentration of labile palladium species inside cells.

Although our current lysine-decaging efficiency on intracellularproteins is still low, this strategy may serve as a useful gain-of-func-tion tool in rescuing the native activity of a target protein that cangenerate an amplifiable signal, such as an enzyme, a transcription

a b

f

0

20

40

60

80

100

120

Cel

l via

bilit

y (%

)

0

5

Cytoso

l

HEK293T

NIH3T

3A54

9HeL

a

Caco-

2CHO

Mem

bran

e

Nucleu

sTot

al

10

15

20

25

30

Pal

ladi

um u

ptak

e(n

g pe

r 10

6 ce

lls)

0

2

4

6

8

10

12

Rel

ativ

e flu

ores

cenc

ein

tens

ity (

a.u.

)

HEK293T

NIH3T3

A549HeLa

Caco-2

CHO

c

Cell type Cell typeCell fraction

170100

5540

35

25

15

Coomassie In-gelfluorescence

His α-tubulin

MW (kDa)

GFP-Y40-Proc-Lys(CuAAC in cell lysates)

Allyl 2Pd 2

Cl 2

-Pd

Allyl 2Pd 2

Cl 2

-Pd

(10 µ

M)

(10 µ

M)

e

Lane: 1 2 3 4

GFP

HN O

O

Treat cells with [Pd]

In-gel fluorescence analysis

Unreacted

Click labelled

Label with Cy3–N3

Unlabelled

POI

HN O

O

POINH2

NH2

POI

HN O

O

POI NN

N

Cy3in cell lysates

d

Products

POI

10 μM allyl2Pd2Cl210 μM Pd(dba)2 0.1% DMSO

10 μM Pd(dba)2

10 μM allyl2Pd2Cl2

10 μM allyl2Pd2Cl2

Control

Peptide sequence Origin Intensity ratio

FSVSGEGEGDAT(ProcK)GK

GFP-Y40-ProcLys 73%

FSVSGEGEGDATK GFP-Y40-Lys 27%

(2) LC-MS-MS analysis of digested peptides

GFP-Y40-ProcLys + allyl2Pd2Cl2 (inside HeLa cells)

Relative fluorescence Decaging yield

69% 31%

Intact protein labelling analysis(1)

Figure 4 | Palladium-mediated propargyl carbamate cleavage within living cells. a, Cytotoxicity analysis of the two identified palladium reagents in six

different cell lines (HeLa, CHO, HEK293T, NIH3T3, Caco-2 and A549). Cells were all treated with palladium compounds for three hours and then cultured for

another 24 hours in fresh media before being subjected to MTS assay. Error bars represent+1 s.d. from three independent experiments. b, Cellular uptake

and distribution of allyl2Pd2Cl2 analysed by ICP-MS. The palladium contents in all three cell fractions significantly increased after palladium treatment. Error

bars represent+1 s.d. from three independent experiments. c, In-cell fluorescence assay of palladium-induced decaging of 5 within different cell lines. Error

bars represent+1 s.d. from three independent experiments. d, Scheme illustrating the utilization of the alkyne moiety from the Proc group to quantify the

depropargylation efficiency on proteins within living cells. POI, protein of interest. e, In-gel fluorescence analysis for determining the depropargylation yield on

the GFP model protein within HeLa cells. Cells expressing GFP-Y40-ProcLys were treated with palladium compounds followed by cell lysis and labelling with

an azide-Cy3 probe via a CuAAC reaction. The relative abundance of GFP-Y40-ProcLys in the cell lysate can be quantified by in-gel fluorescence analysis.

Shown are the in-gel fluorescence and Coomassie staining of SDS-PAGE with the immunoblotting verification for GFP expression (anti-His) and equal loading

(anti-a-tubulin). f, Quantifying the reaction yield of depropargylation on GFP-Y40-ProcLys inside HeLa cells through (1) the protocol shown in d and (2) the

LC-MS-MS analysis of the target peptides after trypsin digestion.

ARTICLES NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry356

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 6: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

regulator or a protein involved in a signal-transduction cascade. Forexample, as a result of their catalytic nature, even a small amount ofenzymes may stimulate a cellular process that requires inhibitionefficiency as high as 90%, or even 95% to ascertain the necessityof an enzyme in a specific signalling pathway. In contrast, thegain-of-function study examines the sufficiency of an enzyme inpromoting a biological event, which may only need an enzyme-acti-vation level as low as 10% to drive the specific phenotype1.Moreover, as our genetic-code expansion method is able to site-specifically replace a critical lysine residue by Proc-Lys on alysine-dependent enzyme, this chemically caged protein resemblesthe active-site mutant of the proteins utilized in the chemicalrescue strategy. Such inactive proteins provide a ‘clean background’whereby the chemical decaging-mediated lysine liberation coulddirectly turn on a distinct signal or phenotype that is specific tothe activated protein.

Palladium-triggered activation of a bacterial phosphothreoninelyase within host cells. To show that our palladium-triggeredlysine depropargylation strategy can effectively modulate thefunction of an enzyme that contains a catalytic lysine, we focusedon OspF, a phosphothreonine lyase known to be secreted intohost cells through a Shigella Type III secretion system38. Onentering the host cells, OspF acts as an epigenetic modulator byirreversibly dephosphorylating mitogen-activated protein kinases(MAPKs), such as phosphorylated Erk (p-Erk), which results inaltered host inflammatory transcriptional responses39. An unusualb-elimination mechanism was employed by OspF to removeirreversibly the phosphate group from phosphothreonine in theconserved Thr-Glu-Tyr (T-E-Y) motif on p-Erk, required forMAPK activity, to generate dehydrobutyrine bearing anunsaturated double bond38 (Fig. 5a). A previous study showed thatboth Lys102 and Lys134 are essential residues for thephosphothreonine lyase activity of OspF towards MAPKs. Inparticular, it was proposed that Lys134 is the key catalytic residuethat can donate a pair of electrons to remove the a-hydrogen ofphosphorylated Thr219 on Erk40. We found that the wild-typeOspF protein (WT-OspF) can effectively dephosphorylate the p-Erk protein in vitro (Fig. 5b, lane 4). Interestingly, when Proc-Lyswas site-specifically incorporated into OspF at position Lys134,the generated OspF variant (OspF-K134-ProcLys) exhibited nodephosphorylation activity on p-Erk in vitro (Fig. 5b, lane 3).Therefore, we confirmed that our Proc group can serve as acaging moiety to mask the activity of a specific lysine residue, andthus the associated enzymatic activity, from its embedded lysine-dependent protein. As expected, the addition of 10 mMallyl2Pd2Cl2 or Pd(dba)2 was able to liberate 1-amine effectivelyfrom the caged Proc-Lys residue on OspF-K134-ProcLys, whichled to restored OspF dephosphorylation activity on p-Erk (Fig. 5b,lanes 1 and 2, Supplementary Method 12).

To prove this is the case in vivo, we expressed OspF-K134-ProcLys in HeLa cells followed by palladium-triggered decaging ofthe embedded Proc-Lys residue. The efficiency of this depropargy-lation reaction inside cells was measured by the aforementionedclick labelling method in combination with in-gel fluorescenceanalysis, which showed that the cleavage of the Proc group onOspF-K134-ProcLys reached a yield of 28% (SupplementaryFig. 19). To demonstrate that this palladium-decaged OspF is enzy-matically functional, we examined the dephosphorylation activity ofOspF-K134-ProcLys before and after palladium treatment.Consistent with the above in vitro observations, replacing Lys134by Proc-Lys rendered OspF incapable of removing the phosphategroup from phosphothreonine on p-Erk in vivo (Fig. 5c, lanes 4and 5). In contrast, treatment by 10 mM Pd(dba)2 or allyl2Pd2Cl2of the OspF-K134-ProcLys-expressing cells restored OspF’s depho-sphorylation activity on p-Erk: apparent dephosphorylation of

p-Erk was detectable by an immunoblotting analysis 60 minutesafter palladium treatment with a maximum dephosphorylationachieved at 180 minutes (Fig. 5c, lanes 1 and 2, and Fig. 5d).Control experiments were performed by adding palladium com-pounds to HeLa cells with no plasmid transfection or bearing theplasmid encoding OspF-K134-TAG in the absence of Proc-Lys.Neither of these cases showed obvious dephosphorylation activityon p-Erk (Fig. 5c, lanes 6–9). In addition, cells treated withseveral other metal ions that may be available in living systemsdid not rescue OspF’s activity (Supplementary Fig. 20).Furthermore, to quantify the enzymatic activity ‘switched on’ byour palladium cleavage reagent from Proc-caged OspF within hostcells, we employed a previously reported p-Erk dual luciferaseassay with its luciferase activity proportional to the intracellularp-Erk level38. This reporter strategy offers a tool to monitor the dropof the intracellular Erk phosphorylation level, and thus the restorednative activity of OspF, by measuring the relative luminescencechange38. First, expression of WT-OspF in HeLa cells led to anear 20-fold decrease of luminescence intensity than did the trans-fection of an empty vector as a negative control (Fig. 5e, columns 5and 6). Whereas the luminescence level for cells that expressedOspF-K134-ProcLys was as high as that of control cells, adding10 mM allyl2Pd2Cl2 led to a significant decrease (�7.5-fold) of lumi-nescence intensity (Fig. 5e, columns 2 and 4). By deducting the rela-tive intracellular levels of WT-OspF and OspF-K134-ProcLys, thisrecovered activity was calculated to be over 80% that of wild-typeOspF (see the detailed calculations in Supplementary Method 16).As additional controls, HeLa cells bearing the plasmids thatexpress OspF-K134-TAG and the PylRS-tRNA pair without Proc-Lys supplementation exhibited a similar high level of luminescencewith and without palladium compounds (Fig. 5e, columns 1 and 3).Taken together, our gain-of-function study, via the chemically con-trolled in situ regeneration of a specific lysine residue on OspF, pro-vides direct evidence, in line with previous mutagenesis andstructural studies, that Lys134 is the key catalytic residue on thisnewly identified phosphothreonine lyase. Moreover, our palla-dium-triggered deprotection of caged OspF led to the restorationof its enzymatic activity to a level close to that of native OspF.

Chemical decaging revealed intracellular relocalization ofOspF-damaged Erk. Further, we applied our chemical decagingstrategy on OspF to study the intracellular localization of Erk onirreversible dephosphorylation by OspF (Fig. 6a). Erk was knownto undergo phosphorylation and dephosphorylation cycles, whichdetermine its localization in the cytoplasm versus the nucleus ofeukaryotic cells41. Moreover, after a nucleus translocation of p-Erkfrom cytoplasm, the accumulation of Erk, even after itsdephosphorylation, has been observed within the nucleus and thisevent, termed ‘nucleus sequestration’, resulted in retention of Erkin the nucleus to various degrees, depending on the source andintensity of extracellular stimulation42. However, the intracellularlocalization of Erk damaged by OspF via permanentdephosphorylation remains elusive. First, we found that GFP-tagged OspF-K134-ProcLys was retained within the nucleus,which indicates that the nucleus localization of OspF wasindependent of its catalytic activity and that the OspF-mediateddephosphorylation reaction occurs within the nucleus (Fig. 6b).Next, the localization of GFP-fused Erk (GFP-Erk) was monitoredbefore and after the palladium-induced decaging of OspF-K134-ProcLys in HeLa cells that coexpress GFP-Erk and OspF-K134-ProcLys. In untreated cells, the overexpressed GFP-Erkaccumulated in the nucleus after stimulating cells with phorbol12-myristate 13-acetate (PMA), consistent with previous report.Rescued OspF, via the palladium-mediated decaging of OspF-K134-ProcLys, caused a gradual loss of nucleus retention of GFP-Erk. After 180 minutes, 12% of the total GFP-Erk pool was

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887 ARTICLES

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry 357

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 7: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

exported to cytoplasm (Fig. 6c,d, Supplementary Fig. 21). Bycontrast, GFP-Erk remained within the nucleus throughout theexperiment for palladium-treated cells without the expression ofWT-OspF or OspF-K134-ProcLys, or for cells without palladiumtreatment (Supplementary Fig. 22). Together, these observationsreveal that the OspF-dephosphorylated Erk was irreversiblyexported from the nucleus to the cytosol, which led to an

accumulation of damaged Erk in the cytoplasm (Fig. 6a). Theimpairment of Erk’s subcellular balance between the nucleus andthe cytoplasm may largely contribute to OspF’s virulence effectswithin host cells. Therefore, our palladium-triggered decagingstrategy was able to turn on sufficient OspF activity to address itspathological roles in modulating Erk’s phosphorylation level andthus the MAPK signalling cascade.

a

Anti OspF

Anti Erk

Anti p-Erk

WT-OspF

OspF-K134-Proc-Lys

p-Erk

10 µM Pd(dba)2

10 µM allyl2Pd2Cl2

+++ ++++

––+ ––––

––– –+++

–+– ––+–

––– +––+

1 2 3 4 5 6 7Lane

d

ERK ERK

Active site

OspF

c

e

Rel

ativ

e lu

cife

rase

act

ivity

Allyl2Pd2Cl2 (10 µM)

pOspF-K134TAG/PylRS

1 mM ProcLys

WT-OspF

Vector

0

5

10

15

20

1 2 3 4 5 6

+

++

– – –

– +

++ – –

+

+

– –

– – – – +–

– – – – + –

Anti OspF

Column

Rel. int.

OspF-K134TAG/PylRS

WT-OspF

1 mM Proc-Lys

10 µM Pd(dba)2

10 µM allyl2Pd2Cl2

Anti OspF

Anti Erk

Anti p-Erk

+ + ++

+ + +–

+ – ––

– + ––

– –– – – –– –+

– + +

– –– ––

+

+

––

1 2 3 4 5 6 7 8 9

+

+

Lane

Anti OspF

Anti p-Erk

Anti Erk

0 10 30 60 120 180 Time (min)Allyl2Pd2Cl2 (10 µM)

OspF-K134-Proc-Lys

1 2 3 4 5 6Lane

b

0.45 1

2 6

HN C C

CH

HN

O

C

H

O CH3

OHN C C

CH

HN

O

C

H3C

O

N

HH

OH

H2N

H

Elimination

Lys102

Lys134

Tyr156

Active site

P

OspF-K134-Proc-Lys

WT-OspF

Dephosphorylation

Pd

T E YP

T E Y*OspF

Figure 5 | Modulating the activity of a bacterial phosphothreonine lyase OspF via palladium-mediated decaging of lysine in vitro and in vivo.

a, Mechanism of the OspF-catalysed irreversible dephosphorylation of its substrate p-Erk. The key catalytic lysine residue 134 on OspF is highlighted in red.

Our palladium-triggered Proc-Lys depropargylation strategy on OspF may modulate its irreversible dephosphorylation activity on p-Erk (shown in the yellow-

shaded box). b, The dephosphorylation assay on p-Erk performed in vitro. Purified WT-OspF can dephosphorylate p-Erk effectively (lane 4), and the

OspF-K134-ProcLys variant protein exhibited no dephosphorylation activity on p-Erk (lane 3). The addition of palladium compounds can convert the Proc-Lys

residue into free Lys and thus restore OspF’s dephosphorylation activity (lanes 1 and 2). c, The p-Erk dephosphorylation assay performed in vivo. In contrast

to WT-OspF, the expressed OspF-K134-ProcLys protein inside the cells exhibited no dephosphorylation activity on p-Erk (lanes 4 and 5). The addition of

palladium compounds in live cells can restore OspF’s dephosphorylation activity (lanes 1 and 2). d, Time-dependent Osp-K134-ProcLys activity restoration in

vivo. e, Dual luciferase assay measures the restoration of relative WT-OspF activity from OspF-K134-ProcLys on palladium treatment in living cells.

The relative luminescence intensity is proportional to the level of intracellular p-Erk. Columns 2 and 4 verify that the addition of allyl2Pd2Cl2 in cells bearing

OspF-K134-ProcLys caused a significant decrease of p-Erk level within HeLa cells. Error bars represent+1 s.d. from three independent experiments.

The relative intensities (Rel. int.) of WT-OspF and OspF-K134-ProcLys in living HeLa cells are shown for comparison.

ARTICLES NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry358

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 8: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

OspF is the prototype of a newly identified OspF family of phos-phothreonine lyases that, as yet, has no homologues in eukaryoticcells. The extremely high specificity of OspF towards MAPKs (Erkand p38) made it an attractive protein-based inhibitor for suchkinases that have no potent small-molecule inhibitors currentlyavailable43. Recent work used OspF as a tool to rewire kinase path-ways in yeast and immune cells, which may provide a new strategy toengineer cells for therapeutic or biotechnological applications43.However, OspF irreversibility dephosphorylates its substrates andthus permanently impairs MAPK signalling pathways, and so thecontrolled activation of this enzyme with spatial and temporal pre-cision is desired. Our chemical decaging strategy may create a‘proenzyme’ form for OspF that would rely on the spatial–temporaladministration of a chemical reagent for selective OspF activationand MAPK targeting (Fig. 6a). The recently developed palladium-immobilized nanoparticles may serve as an excellent proenzymeactivator capable of targeting different subcellular compartmentsor even different cell types. Efforts towards this end are currentlyunderway in our laboratory.

Finally, we demonstrated the general applicability of our strategyin additional cell lines and proteins. We first performed the in situactivation of OspF inside all six cell lines mentioned above, and we

then expanded our strategy to other types of lysine residues on pro-teins, including an essential PTM lysine residue (K372) on thetumour suppressor p5344 and a key lysine residue within the recog-nition motif (Y-P-K-N) on the substrate of SrtD enzyme45

(Supplementary Information).

ConclusionsIn summary, we successfully harnessed a palladium-mediateddeprotection chemistry to control the activation of a given proteinin vitro and in vivo. This was achieved by rational design and catalystscreening for a biocompatible protection group/catalyst pair onfluorogenic small-molecule reporters as well as on intact fluorescentproteins, followed by validating the efficiency and compatibility withlive cells. The Proc moiety served as an effective caging group tomask the activity of a lysine residue from an intact protein,whereas our identified palladium catalysts provided a facileapproach to convert this genetically encoded chemically cagedlysine analogue into free lysine, which led to rescued protein activityunder living conditions. This study represents the first example, toour knowledge, of a transition metal utilized to activate a targetprotein specifically within its native cellular context. Our strategymay be generally applicable to control chemically the liberation of

0.0

0.5

1.0

1.5

2.0

0 min 180 min

F n/F

c

Reaction time (min)

c

d

b

a

OspF-GFPOspF-K134-ProcLys-GFP Erk-GFP

X

Elimination

OspFactive

OspFinactive

ERK

Dephosphorylation

ERK

Export

PhosphorylationImport

(MAPKK)

Reversible

Nucleus

ERKERK

Export

X(MAPKK)

Accumulationin cytosol

Irreversible

Cytosol

This study

OspF secretion

Host cell

OspF:phosphothreonine lyase

NH2HN O

O

OP

OH

Pd

20 μm 20 μm

0 min 60 min 120 min 180 min

20 μm 20 μm20 μm20 μm

Figure 6 | Chemically rescued OspF as a tool to study the intracellular localization of its cognitive substrate Erk. a, A model illustrating that our chemical

decaging strategy on a bacterial secretion effector, OspF, reveals the fate of OspF-damaged MAPK substrates (for example, p-Erk) within host cells. The

palladium-mediated intracellular rescue of the chemically caged OspF shows that this bacterial toxin can cause the impairment of Erk’s nucleus–cytoplasm

localization via irreversible dephosphorylation. b, Imaging of the subcellular distribution of OspF-GFP and OspF-K134-ProcLys-GFP confirmed the overlapped

localization of these two OspF variants. c, Time-lapse imaging of the subcellular distribution of EGFP-Erk after the addition of 10mM allyl2Pd2Cl2 on live HeLa

cells co-expressing OspF-K134-ProcLys. Scale bars, 20mm. d, Normalized Fn /Fc (ratio of nuclear to cytoplasmic EGFP mean fluorescence intensities) at 0

and 180 minutes after allyl2Pd2Cl2 (10 mM) treatment. Error bars indicate+s.d. from ten cells (Supplementary Fig. 21).

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887 ARTICLES

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry 359

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 9: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

key lysine residue(s) from a given protein, and so facilitate intra-cellular manipulation of its lysine-dependent activity. In addition,our work extends the rapidly expanding toolkit of palladium-mediated intracellular cleavage reactions from small molecules toproteins, which may ignite more interest in exploiting transitionmetals for the manipulation of proteins under living conditions46.

Importantly, whereas our palladium catalysts alone did not causethe change in the Erk phosphorylation level in vitro and in vivo, ourresults indicate that ultraviolet irradiation triggered significant p-Erkdephosphorylation when applied to living HeLa cells. Therefore, ourchemical decaging strategy developed here offers a complimentaryapproach based on small molecules to the commonly used photoca-ging methods that rely on photosensitive UAAs for site-specificcontrol of protein activity, particularly when ultraviolet-inducedphotodamage becomes a concern. Furthermore, in comparisonwith small-molecule allosteric activators or the chemical rescue strat-egy, our approach possesses unique features:

(1) As a result of its mechanism-based design, this method is gen-erally applicable to virtually any given protein that containsthe lysine of interest.

(2) The mutant proteins utilized in the chemical rescue strategyusually exhibit low, but detectable, wild-type activity, whichmay interfere with the subsequent activation process to acertain extent, particularly when the protein has an amplifiablesignal (for example, enzymes). In contrast, in our method thecatalytic lysine residue is caged directly, which completelyabolishes the protein activity and thus creates a clean back-ground until the cleavage reagent-mediated decaging.

(3) Sometimes an in vitro activation compound turns out to be aninhibitor for the same enzyme within cells, probably becauseof the interference with protein’s intracellular functions, suchas substrate recruitment1. In contrast, our strategy only createsa point mutation in the active pocket, but does not affectother areas of the protein. This, in conjunction with thesingle-component palladium reagent we used, could largelyavoid such problems.

(4) Instead of non-covalent small-molecule binders that reversiblyswitch the protein between on and off states, our chemical deca-ging strategy creates an irreversible transition on a protein thatpermanently unleashes its activity. This unique mechanismmay largely account for our observations by the proof-of-concept study on OspF that even a modest intracellular decagingefficiency can restore abundant enzymatic activity sufficientto probe a protein’s physiological and pathological rolesinside cells.

A fast-growing list of UAAs34, including various caging ana-logues of the canonical 20 amino acids, have been geneticallyencoded in diverse living cells as well as in multicellular organ-isms13,35. This, in conjunction with the rapidly emerging biofriendlyusage of transition metals in various intracellular chemical conver-sions, may enable a reservoir of deprotection chemistry to controlthe release of amino-acid residues other than lysine. Such a biocom-patible chemical decaging strategy would permit small-moleculecatalysts to modulate selectively the function of an intracellularprotein of interest, particularly in a gain-of-function and site-specific fashion.

MethodsThe chemical synthesis procedures and detailed protocols are all included in theSupplementary Information.

Propargyl carbamate cleavage reaction on purified proteins. Palladium reagents(Sigma) were prepared either as a 10 mM stock in DMSO solution (1:1,000 dilutionwith water solution when used) or directly dissolved in water for Pd(NO3)2 andNa2PdCl4. Phosphine ligands (P(PhSO3Na)3, JK Chemicals) were prepared as

20 mM stock solutions in pure water. All these materials were prepared just beforeuse. For in vitro propargyl carbamate cleavage on purified proteins, recombinantGFP protein bearing the site-specifically incorporated Proc-Lys (all at 10 mM finalconcentration in 1 × DPBS buffer, pH 7.4) were purified and incubated with10 equiv. palladium compounds (100 mM final concentration). The finalconcentration for the phosphine ligands used was 200 mM. The solution wasincubated at room temperature for one hour and then quenched by adding 10 ml3-mercaptopropanoic acid solution (1%) per 100 ml reaction system. The quenchedreaction was left at room temperature for 15 minutes and then desalted through aBio-spin 6 column (Bio-Rad) before being analysed by fluorescence measurement,native fluorescence gel and LC-ESI-MS. The native gels were either stained withCoomassie blue or directly scanned with a ChemiDo XRSþ System (Bio-Rad) tovisualize in-gel fluorescence derived from native GFP.

Expression and depropargylation of Proc-Lys that contains proteins inmammalian cells. For Proc-Lys incorporation, a plasmid that encoded the targetprotein carrying an in-frame amber mutation was co-transfected with a plasmidexpressing the PylRS/tRNACUA

Pyl pair into cells via X-tremeGENE HP (Roche) inDMEM (10% FBS) supplemented with 1 mM Proc-Lys. Cells were allowed to growfor an additional 24 hours to express the desired protein bearing a site-specificallyincorporated Proc-Lys residue. For depropargylation on proteins, cells thatharboured the Proc-Lys-incorporated proteins were first cultured in DMEM withoutProc-Lys and FBS for 180 minutes, followed by treatment with palladium reagents(10 mM) in fresh DMEM for another 180 minutes to allow the propargyl carbamatecleavage reaction to proceed inside mammalian cells.

Quantifying propargyl carbamate cleavage efficiency on intracellular proteins.The Proc-Lys-bearing protein within cells before and after the palladium-mediateddepropargylation were labelled by an azide-Cy3 probe (200 mM final concentration)via the Cu(I)/BTTAA (BTTAA¼ 2-(4-((bis((1-tert-butyl-1H-1,2,3-triazol-4-yl)methyl)amino)methyl)-1H-1,2,3-triazol-1-yl)acetic acid) (100 mM finalconcentration, Cu/BTTAA¼ 1:2) mediated CuAAC reaction in cell lysate(2 mg ml21 final concentration, quantified by bicinchoninic acid assay). Thelabelling reaction was allowed to proceed for one hour at 30 8C before beingquenched by BCS and 5 × SDS sample buffer. The labelled lysates were thensubjected to SDS–PAGE and in-gel fluorescence analysis (Typhoon-FLA9500, GE)to quantify the propargyl carbamate cleavage efficiency on proteins. In a separateapproach, the target protein was purified or immunoprecipitated from cells andsubjected to in-gel trypsin digestion. The deprotection efficiency was quantified bycomparing the Proc-group protected and deprotected peptides from LC-MS-MSanalysis.

Palladium-mediated depropargylation of OspF in mammalian cells. HeLa cellswere seeded into a 24-well corning plate and left to grow to approximately 70%confluence in DMEM that contained 10% FBS and penicillin/streptomycin.Co-transfection was performed by plasmids that encode pcDNA4-HA-flag-WT-OspFor pcDNA4-HA-flag-OspF-134TAG with a plasmid expressing thePylRS/tRNACUA

Pyl pair using X-tremeGENE HP (Roche) in DMEM (10% FBS)supplemented with and without 1 mM Proc-Lys. Cells were allowed to grow for anadditional 24 hours followed by incubation in DMEM without Proc-Lys and FBS for180 minutes. Palladium reagents (10 mM) were used to treat the cells in freshDMEM for 180 minutes, after which the palladium compounds were removed bywashing and the cells were incubated in fresh DMEM that contained 0.01% PMA for30 minutes to activate p-Erk. Cells that contained the activated p-Erk were then lysedusing 5 × SDS sample buffer and boiled at 95 8C for 30 minutes, followed by theseparation of soluble fractions from cell debris and SDS–PAGE analysis. Westernblotting analysis was carried out to detect the full-length OspF protein as well as thedephosphorylation level of p-Erk by using antibodies against OspF, p-Erk1/2 or Erk1/2.

Dual luciferase assay to monitor the phosphorylation level of Erk. The in vivodual luciferase assay on p-Erk was performed following the manufacturer’sinstructions (Promega) and a previous report38. Plasmids that encode Gal4-Elk,Gal4-luc and pRL-TK were transfected in combination with plasmids used in theabove western blotting analysis. Cells after transfection and palladium treatmentwere incubated in fresh medium for another 12 hours before the relative p-Erk levelswere analysed by determining the luciferase activity using a microplate reader(Bio-Tek, Synergy H4).

Live-cell fluorescence imaging. HeLa cells were seeded into a 24-well corning plateand allowed to grow to approximately 90% confluence in DMEM (10% FBS) thatcontained penicillin/streptomycin. Transfection was carried out using the sameprotocol as in the western blotting analysis. After transfection, cells were grown for18 hours and transferred onto an eight-well Lab-Tek chambered coverglass foranother eight hours in DMEM and in the absence of Proc-Lys. The p-Erk level incells were first elevated by PMA (0.01%) followed by treatment with palladiumreagents (10 mM) for three hours and fluorescence imaging by a Zeiss LSM 700 laserscanning microscope with a Plan Apochromat ×63 or ×40 oil-immersion objectivein a scan zoom (averaging 16). The mean fluorescence intensities of nucleus (Fn) andcytoplasm (Fc) were quantified using ZEN software to enable the Fn/Fc ratio to bedetermined according to the formula Fn/Fc¼ (Fn – Fb)/(Fc – Fb), where Fb is themean background fluorescence intensity.

ARTICLES NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry360

© 2014 Macmillan Publishers Limited. All rights reserved.

Page 10: Palladium-triggered deprotection chemistry for protein ... · Palladium-triggered deprotection chemistry for protein activation in living cells Jie Li1, Juntao Yu2,JingyiZhao1,JieWang2,

Received 14 August 2013; accepted 5 February 2014;published online 16 March 2014

References1. Zorn, J. A. & Wells, J. A. Turning enzymes ON with small molecules. Nature

Chem. Biol. 6, 179–188 (2010).2. Bishop, A. & Chen, V. Brought to life: targeted activation of enzyme function

with small molecules. J. Chem. Biol., 2, 1–9 (2009).3. Riggsbee, C. W. & Deiters, A. Recent advances in the photochemical control of

protein function. Trends Biotechnol. 28, 468–475 (2010).4. Banaszynski, L. A., Chen, L-C., Maynard-Smith, L. A., Ooi, A. G. L. &

Wandless, T. J. A rapid, reversible, and tunable method to regulate proteinfunction in living cells using synthetic small molecules. Cell 126,995–1004 (2006).

5. Ghosh M. et al. Cofilin promotes actin polymerization and defines the directionof cell motility. Science 304, 743–746 (2004).

6. Wolan, D. W., Zorn, J. A., Gray, D. C. & Wells, J. A. Small-molecule activators ofa proenzyme. Science 326, 853–858 (2009).

7. Erster O., Eisenstein, M. & Liscovitch, M. Ligand interaction scan: a generalmethod for engineering ligand-sensitive protein alleles. Nature Methods 4,393–395 (2007).

8. Qiao, Y., Molina, H., Pandey, A., Zhang. J. & Cole, P. A. Chemical rescue of amutant enzyme in living cells. Science 311, 1293–1297 (2006).

9. Pratt, M. R, Schwartz, E. C. & Muir, T. W. Small-molecule-mediated rescue ofprotein function by an inducible proteolytic shunt. Proc. Natl Acad. Sci. USA104, 11209–11214 (2007).

10. Li, H., Hah, J-M. & Lawrence, D. S. Light-mediated liberation of enzymaticactivity: ‘small molecule’ caged protein equivalents. J. Am. Chem. Soc. 130,10474–10475 (2008).

11. Zou, K., Cheley, S., Givens, R. S. & Bayley, H. Catalytic subunit of protein kinasea caged at the activating phosphothreonine. J. Am. Chem. Soc. 124,8220–8229 (2002).

12. Chen, P. R. et al. A facile system for encoding unnatural amino acids inmammalian cells. Angew. Chem. Int. Ed. 48, 4052–4055 (2009).

13. Davis, L. & Chin, J. W. Designer proteins: applications of genetic code expansionin cell biology. Nature Rev. Mol. Cell Biol. 13, 168–182 (2012).

14. Arbely, E., Torres-Kolbus, J., Deiters, A. & Chin, J. W. Photocontrol of tyrosinephosphorylation in mammalian cells via genetic encoding of photocagedtyrosine. J. Am. Chem. Soc. 134, 11912–11915 (2012).

15. Oksvold, M. P., Huitfeldt, H. S., Østvold, A. C. & Skarpen, E. UV inducestyrosine kinase-independent internalisation and endosome arrest of the EGFreceptor. J. Cell Sci. 115, 793–803 (2002).

16. Walsh, C. T. Posttranslational Modification of Proteins: Expanding Nature’sInventory 1st edn (Roberts and Company, 2005).

17. Carrera, A. C., Alexandrov, K. & Roberts, T. M. The conserved lysine of thecatalytic domain of protein kinases is actively involved in the phosphotransferreaction and not required for anchoring ATP. Proc. Natl Acad. Sci. USA 90,442–446 (1993).

18. Streu, C. & Meggers, E. Ruthenium-induced allylcarbamate cleavage in livingcells. Angew. Chem. Int. Ed. 45, 5645–5648 (2006).

19. Garner, A. L., Song, F. & Koide, K. Enhancement of a catalysis-basedfluorometric detection method for palladium through rational fine-tuning of thepalladium species. J. Am. Chem. Soc. 131, 5163–5171 (2009).

20. Isidro-Llobet, A., Alvarez, M. & Albericio, F. Amino acid-protecting groups.Chem. Rev. 109, 2455–2504 (2009).

21. Ai, H-W., Lee, J. W. & Schultz, P. G. A method to site-specifically introducemethyllysine into proteins in E. coli. Chem. Commun. 46, 5506–5508 (2010).

22. Santra, M., Ko, S.-K., Shin, I & Ahn, K. H. Fluorescent detection ofpalladium species with an O-propargylated fluorescein. Chem. Commun. 46,3964–3966 (2010).

23. Yusop, R. M., Unciti-Broceta, A., Johansson, E. M. V., Sanchez-Martın, R. M. &Bradley, M. Palladium-mediated intracellular chemistry. Nature Chem. 3,239–243 (2011).

24. Sasmal, P. K., Streu, C. N. & Meggers, E. Metal complex catalysis in livingbiological systems. Chem. Commun. 49, 1581–1587 (2013).

25. Li. J. et al. Ligand-free palladium-mediated site-specific protein labeling insidegram-negative bacterial pathogens. J. Am. Chem. Soc. 135, 7330–7338 (2013).

26. Chalker, J. M., Wood, C. S. C. & Davis, B. G. A convenient catalyst for aqueousand protein Suzuki–Miyaura cross-coupling. J. Am. Chem. Soc. 131,16346–16347 (2009).

27. Fors, B. P., Krattiger, P., Strieter, E. & Buchwald, S. L. Water-mediated catalystpreactivation: an efficient protocol for C–N cross-coupling reactions. Org. Lett.10, 3505–3508 (2008).

28. Ito, T., Takahashi, Y. & Ishii Y. Convenient preparation of novel palladium-p-olefin complexes from bis(dibenzylideneacetone)palladium(0). Chem. Commun.629a–629a (1972).

29. Wei, C.S. et al. The impact of palladium(II) reduction pathways on thestructure and activity of palladium(0) catalysts. Angew. Chem. Int. Ed. 52,5822–5826 (2013).

30. Li, N., Lim, R. K. V., Edwardraja, S. & Lin, Q. Copper-free Sonogashira cross-coupling for functionalization of alkyne-encoded proteins in aqueous mediumand in bacterial cells. J. Am. Chem. Soc. 133, 15316–15319 (2011).

31. Rambabu, D., Bhavani, S., Swamy, N. K., Basaveswara Rao, M. V. & Pal, M.Pd/C-mediated depropargylation of propargyl ethers/amines in water.Tetrahedron Lett. 54, 1169–1173 (2013).

32. Hao, B. et al. A new UAG-encoded residue in the structure of a methanogenmethyltransferase. Science 296, 1462–1466 (2002).

33. Mukai T et al. Adding L-lysine derivatives to the genetic code of mammalian cellswith engineered pyrrolysyl-tRNA synthetases. Biochem. Biophys. Res. Commun.371, 818–822 (2008).

34. Liu, C. C. & Schultz, P. G. Adding new chemistries to the genetic code. Annu.Rev. Biochem. 79, 413–444 (2010).

35. Bianco, A., Townsley, F. M., Greiss, S., Lang, K. & Chin, J. W. Expanding thegenetic code of Drosophila melanogaster. Nature Chem. Biol. 8, 748–750 (2012).

36. Nguyen, D. P. et al. Genetic encoding and labeling of aliphatic azides and alkynesin recombinant proteins via a pyrrolysyl-tRNA synthetase/tRNACUA pair andclick chemistry. J. Am. Chem. Soc. 131, 8720–8721 (2009).

37. Kennedy, D. C. et al. Cellular consequences of copper complexes used to catalyzebioorthogonal click reactions. J. Am. Chem. Soc. 133, 17993–18001 (2011).

38. Li, H. et al. The phosphothreonine lyase activity of a bacterial Type III effectorfamily. Science 315, 1000–1003 (2007).

39. Arbibe, L. et al. An injected bacterial effector targets chromatin access fortranscription factor NF-kB to alter transcription of host genes involved inimmune responses. Nature Immunol. 8, 47–56 (2007).

40. Zhu, Y. et al. Structural insights into the enzymatic mechanism of the pathogenicMAPK phosphothreonine lyase. Mol. Cell 28, 899–913 (2007).

41. Lidke, D. S. et al. ERK nuclear translocation is dimerization-independent butcontrolled by the rate of phosphorylation. J. Biol. Chem. 285, 3092–3102 (2010).

42. Volmat, V., Camps, M., Arkinstall, S., Pouyssegur, J. & Lenormand, P. Thenucleus, a site for signal termination by sequestration and inactivation ofp42/p44 MAP kinases. J. Cell Sci. 114, 3433–3443 (2001).

43. Wei, P. et al. Bacterial virulence proteins as tools to rewire kinase pathways inyeast and immune cells. Nature 488, 384–388 (2012).

44. Chuikov, S. et al. Regulation of p53 activity through lysine methylation. Nature432, 353–360 (2004).

45. Hendrickx, A. P. A., Budzik, J. M., Oh, S-Y. & Schneewind, O. Architects at thebacterial surface – sortases and the assembly of pili with isopeptide bonds.Nature Rev. Micro. 9, 166–176 (2011).

46. Li, J. & Chen, P. R. Moving Pd-mediated protein cross coupling to living systems.Chembiochem 13, 1728–1731 (2012).

AcknowledgementsThis work was supported by research grants from National Natural Science Foundation ofChina (21225206 and 91313301) and the National Key Basic Research Foundation of China(2010CB912302). We thank F. Shao and O. Schneewind for the donation of plasmids.

Author contributionsP.R.C. conceived and designed the experiments. J.L., J.Y., J.Z., J.W., S.Z., S.L., L.C., M.Y., S.J.and X.Z. performed the experiments. P.R.C., J.L., J.Y., J.Z. and J.W. analysed the data. P.R.C,J.L. and J.Y. prepared the figures and co-wrote the paper, with input from all the authors.

Additional informationSupplementary information is available in the online version of the paper. Reprints andpermissions information is available online at www.nature.com/reprints. Correspondence andrequests for materials should be addressed to P.R.C.

Competing financial interestsThe authors declare no competing financial interests.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1887 ARTICLES

NATURE CHEMISTRY | VOL 6 | APRIL 2014 | www.nature.com/naturechemistry 361

© 2014 Macmillan Publishers Limited. All rights reserved.


Recommended