1196 VOLUME 35 NUMBER 12 DECEMBER 2017 nature biotechnology

l e t t e r s

Although advances in protein labeling methods have made it possible to measure the proteome of mixed cell populations, it has not been possible to isolate cell-type-specific proteomes in vivo. This is because the existing methods for metabolic protein labeling in vivo access all cell types. We report the development of a transgenic mouse line where Cre-recombinase-induced expression of a mutant methionyl-tRNA synthetase (L274G) enables the cell-type-specific labeling of nascent proteins with a non-canonical amino-acid and click chemistry. Using immunoblotting, imaging and mass spectrometry, we use our transgenic mouse to label and analyze proteins in excitatory principal neurons and Purkinje neurons in vitro (brain slices) and in vivo. We discover more than 200 proteins that are differentially regulated in hippocampal excitatory neurons by exposing mice to an environment with enriched sensory cues. Our approach can be used to isolate, analyze and quantitate cell-type-specific proteomes and their dynamics in healthy and diseased tissues.

Measuring the proteomes of specific cell types and analyzing their similarities and differences is essential for understanding both the function and dysfunction of multi-cellular organisms. Bio-orthogonal strategies based on metabolic precursors allow labeling of proteins or the introduction of measurable protein modifications1–3. Because the incorporation of precursors into proteins requires processing by endogenous cell machinery, bioorthogonal labeling strategies are par-ticularly amenable to genetic control. For example, the expression of a methionyl-tRNA synthetase (MetRS) with an expanded amino-acid binding site (MetRS L274G) enables the methionine tRNA to be charged with the methionine surrogate azidonorleucine (ANL)4. By controlling transcription of a gene encoding MetRS L274G with cell-type-specific promoters it is possible to direct cell-type-specific incor-poration of ANL into nascent polypeptides, which can then be tagged by a CLICK reaction and imaged or isolated by affinity purification5. Adding to recent advances in invertebrate model organisms6,7, we implemented and developed MetRS L274G in mouse, the most ver-satile mammalian genetic model organism. We engineered a mouse

line in which the conditional expression of MetRS L274G is under the control of Cre recombinase. When crossed with animals expressing Cre under the control of a cell-type-specific promoter, this approach enables the metabolic labeling of proteins in any cell type for which a Cre-driver mouse line exists. Using our method we labeled, visual-ized and identified nascent proteins in excitatory principal neurons (hippocampus) and Purkinje neurons (cerebellum).

To enable cell-type-selective non-canonical amino-acid tagging in mammals, we cloned and mutated the mouse MetRS to introduce a L274G point mutation at the position analogous to the L13 position in the bacterial protein (Fig. 1a)4,8. Mouse MetRS L274G (MmMetRS-L274G, hereafter referred to as MetRS*) was co-expressed with green fluorescent protein (GFP) from a GFP-2A-MetRS* construct (Supplementary Figs. 1a and 2g). To test conditional protein meta-bolic labeling in mammalian cells, we performed bio-orthogonal non-canonical amino-acid tagging (BONCAT)2 in COS7 or HEK cells transfected with GFP-2A-MetRS* after metabolic labeling with ANL. As expected, ANL incorporation into protein was detected only in cells expressing MetRS*, and could be detected after ANL incuba-tion periods as short as 20 min (Supplementary Figs. 1b and 2a–c). Proteins in MetRS*-expressing (HeLa) cells treated with ANL exhib-ited turnover rates indistinguishable from controls (Supplementary Fig. 2d–f).

To enable in vivo cell-type-specific nascent proteome labeling, a transgenic mouse line was generated by inserting a floxed-STOP version of the GFP-2A-MetRS* construct for conditional expression under a CAG promoter9 in the ROSA26 (R26) locus10 (Supplementary Fig. 1a). To assess conditional expression in specific neuronal types in vivo, R26-MetRS* mice were crossed with two different Cre-driver lines: CaMK2a-Cre and glutamic-acid decarboxylase 2 (GAD2-Cre), to induce GFP-2A-MetRS* expression in glutamatergic (excitatory) neu-rons11 and GABAergic (inhibitory) neurons12 (Online Methods and Supplementary Fig. 1d,e). To visualize GFP-2A-MetRS* expression, whole brain slices were immunolabeled with an anti-GFP antibody. In CaMK2a-Cre::R26-MetRS* mice, MetRS* expression was prima-rily restricted to hippocampal CA1 pyramidal (excitatory) neurons, dentate granule (excitatory) cells and to neurons in the striatum and

Cell-type-specific metabolic labeling of nascent proteomes in vivoBeatriz Alvarez-Castelao1, Christoph T Schanzenbächer1,2,6, Cyril Hanus1,5,6, Caspar Glock1, Susanne tom Dieck1, Aline R Dörrbaum1,2, Ina Bartnik1, Belquis Nassim-Assir1, Elena Ciirdaeva1, Anke Mueller3, Daniela C Dieterich3, David A Tirrell4, Julian D Langer1,2 & Erin M Schuman1

1Max Planck Institute for Brain Research, Frankfurt, Germany. 2Max Planck Institute of Biophysics, Frankfurt, Germany. 3Institute for Pharmacology and Toxicology, Otto von Guericke University, Magdeburg, Germany; Leibniz Institute for Neurobiology, Magdeburg, Germany; and Center for Behavioral Brain Sciences, Magdeburg, Germany. 4Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California, USA. 5Present address: Center for Psychiatry and Neurosciences, Inserm 894, Paris, France. 6These authors contributed equally to this work. Correspondence should be addressed to E.M.S. (erin.schuman@brain.mpg.de).

Received 12 December 2016; accepted 19 October 2017; published online 6 November 2017; doi:10.1038/nbt.4016

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

nature biotechnology VOLUME 35 NUMBER 12 DECEMBER 2017 1197

l e t t e r s

cortex (Supplementary Figs. 1d and 3a,b). In contrast, in GAD2-Cre::R26-MetRS* mice, MetRS* expression was most prominent in Purkinje (inhibitory) neurons of the cerebellum (Supplementary Figs. 1d and 3c,d). To examine the potential for labeling and visual-izing nascent proteins in a genetically targeted cell type, hippocampal slices from CaMK2a-Cre::R26-MetRS* or wild-type mice were pre-pared and incubated with artificial cerebrospinal fluid (ACSF), sup-plemented with ANL or methionine and then processed for fluorescent non-canonical amino-acid tagging (FUNCAT), enabling visualization

of newly synthesized proteins in situ13 (Fig. 1b). As expected, MetRS* expression in CA1 pyramidal neurons coupled with ANL treatment resulted in a strong FUNCAT signal; newly synthesized proteins were clearly visible in neuronal somata and in dendrites (Fig. 1c), whereas no signal was detectable in other cell types (e.g., astrocytes) or in Met-treated slices (Supplementary Fig. 4a). Similar results were obtained from MetRS* brain slices transduced with a Cre-containing virus (Supplementary Fig. 4c). Neither expression of the MetRS* transgene nor exposure of hippocampal slices to ANL affected synaptic transmis-sion in the hippocampal slice (Supplementary Fig. 4b).

Using a recently described technique14 we were also able to label newly synthesized proteins of interest, including the synaptic proteins Drebrin (Dbn) and a glutamate receptor (GluA1) as well as the nuclear protein Lamin, in cultured MetRS*-expressing hippocampal neurons treated with ANL (Fig. 1d). Nascent Dbn and GluA1 were clearly visible in both the soma and dendrites of MetRS*-expressing neurons (but not neighboring neurons), whereas the nascent Lamin was confined to the somatic compartment (Fig. 1d). Besides the visualization of nascent pro-teins of interest in cell types of interest, these data show that labeling with ANL does not perturb the normal localization of labeled proteins.

To label newly synthesized proteins in vivo, CaMK2a-Cre::R26-MetRS* (or control WT) mice were supplied with drinking water supplemented with ANL (16–60 mM) and maltose (to increase water intake) (Fig. 2a). Different amounts of ANL and labeling periods were tested (Supplementary Fig. 5a,b); consumption of ANL had no apparent behavioral effect in WT or MetRS* mice (Supplementary Fig. 5c,d). After 21 d of ANL consumption, the hippocampus was dissected and a click reaction with a biotin alkyne tag was per-formed in order to visualize or identify the nascent proteome in excitatory neurons (Fig. 2a). Nascent labeled proteins, detected using FUNCAT13, were clearly visible in hippocampal neurons express-ing MetRS* (Fig. 2b). Western blot analysis revealed an abundance of biotinylated nascent proteins, spanning all molecular weights, in hippocampal tissue from the CaMK2a-Cre::R26-MetRS* and much lower background levels of biotinylated proteins obtained from the WT mice (Fig. 2b and Supplementary Fig. 5a,b). For protein puri-fication, metabolically labeled proteins were clicked with a cleav-able biotin-alkyne tag, affinity-purified and then selectively eluted by the cleavage and reduction of the alkyne. Eluted proteins were digested and identified using liquid-chromatography-coupled tan-dem mass spectrometry (Supplementary Fig. 5e). To identify the CaMK2a hippocampal proteome, we analyzed proteins that were detected exclusively or enriched in CaMK2a-Cre::R26-MetRS* compared to the corresponding WT control experiments. To this end, we filtered for proteins exclusively detected in CaMK2a samples (in three of six replicates and in none of the WT samples) and for proteins more than threefold enriched in at least three biological CaMK2a replicates (Supplementary Fig. 6). We detected a total of 2,384 proteins (1,110 unique proteins and 1,274 enriched proteins; Fig. 3c,d, Supplementary Table 1 and Supplementary Fig. 7a). To address whether the ANL-labeled proteome is biased, we compared the protein lengths and methionine content of the CaMK2a-Cre::R26-MetRS* proteome to other proteomes and found no significant difference (Supplementary Fig. 7b,c).

To understand the functional nature of the CaMK2a hippoc-ampal proteome, we performed a gene ontology (GO) analysis for cellular components and processes and found a significant enrich-ment of proteins represented by neuronal components (Fig. 2e and Supplementary Table 2). Importantly, terms associated with syn-aptic transmission and synaptic plasticity were among the most significantly enriched GO groups, thus supporting the fidelity of the

Dbn GluA1 Lamin

FU

NC

AT

-PLA

GF

P-I

F

d

FU

NC

AT

GF

P-I

F

CA1CA3

DG

CA1

CA3

FUNCAT (CA1)

DG

c

bCaMK2a-Cre::R26-MetRS*

vs WT

Brain

Hippocampusslice

Metabolic labeling (3 h)

ANL

FUNCAT

a

O

S

OH

H2NO

OH

H2N

L274

MetRS(WT)

Met

AAAA

N

Ribosome

Protein tRNA

ANL

N+N-

NG274

MetRS*(L274G)

AAAA

N N3N3

N3

Figure 1 Genetically targeted protein labeling. (a) A single mutation (L274G) in the amino acid binding site of methionyl-tRNA synthetase (MetRS) enables the loading of the methionine surrogate azidonorleucine (ANL) onto methionine tRNA, and hence the bioorthogonal tagging of nascent proteins synthesized in the cells that express MetRS*. (b) Workflow for the preparation of acute hippocampal sections from CaMK2a-Cre::R26-MetRS* or WT animals and incubation for 4 h with ANL ex vivo. (c) GFP expression and nascent protein labeling in acute hippocampal slices from CaMK2a-Cre::R26-MetRS* mice incubated with ANL for 4 h. CA3, Cornu Ammonis 3; CA1, Cornu Ammonis 1; DG, dentate gyrus. Note the FUNCAT signal in the apical dendrites of CA1 pyramidal cells (arrow). Scale bars, 100 µM, 50 µM and 50 µM, from left to right. (d) Cultured hippocampal neurons expressing GFP-P2A-R26-MetRS* were labeled with ANL for 1 h, and the subcellular localization of newly synthesized Drebrin, GluA1 and Lamin proteins was evaluated using FUNCAT-PLA. Scale bar, 25 µM See Supplementary Figures 1–4.

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

1198 VOLUME 35 NUMBER 12 DECEMBER 2017 nature biotechnology

l e t t e r s

cell-type-specific labeling of excitatory neurons. In addition, within the excitatory neuron proteome we detected several key proteins, the mutations of which give rise to important neurodevelopmental or neurodegenerative disorders (Fig. 2f). To further confirm the cell specificity of the CaMK2a hippocampal proteome, we focused on cell-type-specific markers defined by proteomic and transcriptomic studies. Compared to a whole hippocampal proteome, our data set was significantly de-enriched for glial cell markers and enriched for general neuronal and CA1-pyramidal neuronal markers (P < 0.05, Fig. 2g and Supplementary Table 3).

To examine the versatility of our method, we analyzed the cellular identity of the proteome obtained from the cerebellum of GAD2::R26-MetRS* mice, where (inhibitory) Purkinje cells are predominantly labeled (Supplementary Figs. 6, 7 and 8a). A comparison with pro-teins labeled in control WT mice yielded 1,687 proteins in the GAD2 cerebellar proteome (Supplementary Fig. 8b,c and Supplementary Table 4). A GO analysis revealed a significant enrichment of pro-teins and terms related to metabolism (Supplementary Fig. 8d and Supplementary Table 5). This finding is in good agreement with the special metabolic requirements observed in Purkinje cells15. As above, an analysis of cell-type-specific markers revealed a de-enrichment for glial markers and enrichment for neuronal markers and Purkinje neu-ron markers (Supplementary Fig. 8e and Supplementary Table 3).

We next compared the hippocampal excitatory to the cerebellar inhibitory proteome and discovered 95 and 70 ‘unique’ proteins in each proteome (Fig. 3a, Supplementary Fig. 6 and Supplementary Table 6). By comparing the peptide intensities of the shared (1,382) proteins we discovered 180 and 188 proteins that were significantly enriched in the hippocampal excitatory or cerebellar inhibitory pro-teomes, respectively, likely reflecting important functional differences

in the cell types (permutation-based false discovery rate of 2%, S0=0.1; Fig. 3b and Supplementary Table 6). Among the most enriched pro-teins in the excitatory hippocampal proteome, we found important synaptic proteins such as Dlg4 (PSD-95; ref. 16), a multi-functional scaffolding molecule, and Nckap1, a regulator of the actin cytoskele-ton17. Consistent with the established role of the hippocampal excitatory neurons in learning, we also found enriched proteins related to synaptic plasticity and learning such as Ncdn (neurochondrin18). Intriguingly, we also identified several disease-related proteins such Cyfip2, a protein related to Fragile X syndrome19, and Dync1h1, a protein associated with axonal transport implicated in Charcot–Marie–Tooth disease20. Among the most upregulated proteins in the cerebellar inhibitory proteome was calbindin-1, a canonical Purkinje cell marker and Itpr1, a regulator of calcium channel activity linked to cerebellar Ataxias21.

We carried out a hierarchical clustering of all proteins expressed in both cell types, sorting proteins by expression levels (Fig. 3c). As expected, this showed a consistent pattern of expression within each cell type, but also revealed the two proteomes were distinguishable on the basis of differential expression of proteins. Lastly, in order to examine whether the proteomic data obtained were sufficient to separate distinct tissues and cell types, we conducted a principal component analysis (PCA) including comprehensive hippocampal, cerebellar and glial proteomes (Supplementary Figs. 6 and 7a). The PCA (Fig. 3d) illustrates that the CaMK2a hippocampal proteome and the Gad2 proteome were clearly separated, and the proximity of each to its tissue of origin proteome was also apparent (Fig. 3d). Taken together, these data indicate that this methodology allows for the identification of cell-type-specific proteomes labeled in vivo in complex tissues without mechanical dissociation and isolation of specific cell types.

CaMK2a-Cre::R26-MetRS*vs WT

ANLa

Proteins

MS

m/z

BONCAT

FUNCAT

b WT

KDa

25

250

20

25

50

37

150

100

75

GFP

GFP

cP

rote

in in

tens

ities

in M

etR

S*

(log 2

scal

e)

Protein intensities in wt(log2 scale)

12 18 23 29 34 4012

18

23

29

34

40Enriched proteins

d CaMK2a proteome(unique + enriched)

2,384 206 10

WT

CaMK2a::MetRS*

e

0 1 2 3

NAD metabolic processNegative regulation of dendrite development

Regulation of postsynaptic membrane potentialExcitatory postsynaptic potential

Sodium ion transmembrane transport

Regulation of synaptic transmission, glutamatergicRegulation of dendritic spine morphogenesis

Regulation of long-term neuronal synaptic plasticityRegulation of synaptic vesicle cycle

Regulation of adherens junction organizationG-protein coupled receptor signaling pathway, coupled to

NADH metabolic process

adenylate cyclase-modulating G-protein coupled receptor

Response to ammonium ionCell-cell recognition

Negative regulation of dendrite morphogenesisHydrogen ion transmembrane transport

Motor neuron axon guidance

Regulation of synaptic vesicle exocytosisRegulation of peptidyl-threonine phosphorylation

Protein localization to synapse

Negative regulation of dendritic spine developmentNegative regulation of adherens junction organization

4 5

Enrichment–log10 P-value

Gene ontology(top 25 cellular processes)

Cell recognition

Action potential

6

f

App AlzheimerGrm5 Autism

Huntington

Autism

HttPrnp Creutzfeldt-Jakob

Shank2Snca Parkinson

Disease-related proteins

Cell-typemarker enrichment

g

****

****

****

Glia

Neu

rons

CA

1 py

ram

idal

cel

ls

Protein RNA

Enr

ichm

ent c

ompa

red

to w

hole

hip

poca

mpa

l pro

teom

e

–0.3

0.3

0.2

0.1

0

–0.1

–0.2

0.4

–0.4

–0.5

Camk2MetRS* FUNCAT

BONCAT

Brain

Slice

Figure 2 Cell-type-specific proteomics in vivo: the hippocampal excitatory neuron proteome. (a) CaMK2a-Cre::R26-MetRS* mice were provided with ANL in their drinking water and their brains were collected and processed for FUNCAT (Supplementary Fig. 5) or BONCAT and protein identification by mass spectrometry (MS). (b) FUNCAT of in vivo-labeled proteins in MetRS* mice (left panel) and BONCAT in WT or MetRS* mice that were provided with ANL (60 mM) for 21 d. Scale bar, 20 µM. (c,d) Labeled (biotinylated) proteins were affinity purified, digested and identified by peptide fingerprinting. Scatter plot showing the increased levels (peptide intensities) of proteins found in both MetRS*and WT mice (n = 6 biological replicates for each group) (c). The hippocampal excitatory neuron proteome was obtained by the union of proteins unique to or markedly enriched (over threefold difference from WT) in MetRS* mice (d). (e) Gene ontology analysis. Top 25 cellular processes enriched in the hippocampal excitatory neuron proteome (−log [P values]). Note the presence of multiple terms related to synaptic transmission and plasticity (bold). (f) Example proteins of particular clinical relevance found in the hippocampal excitatory proteome. (g) Relative abundance of previously identified brain-cell-type markers in hippocampal excitatory neuron proteome as compared to the hippocampal proteome. In the left bar, note the de-enrichment for glial markers and to the right, the enrichment for neuronal and CA1 pyramidal neuron markers. See Supplementary Figures 5–7.

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

nature biotechnology VOLUME 35 NUMBER 12 DECEMBER 2017 1199

l e t t e r s

External stimuli lead to changes in cellular proteomes that rep-resent responses or adaptation. To examine whether our method can also detect proteome dynamics, we exposed CaMK2a-Cre::R26-MetRS* mice to an enriched sensory environment (Fig. 4a), which has been shown to modify brain circuits, changing the density of synapses and increasing the volume of the cerebral cortex22. We compared the CaMK2a hippocampal proteomes of mice housed in an enriched environment (EE) to mice that were housed in stand-ard cages (SC) for 21 d. As expected, the two proteomes (2,384 and 2,365 for SC and EE, respectively) (Supplementary Fig. 6 and Supplementary Tables 1 and 7) were largely overlapping (87.2% of

the EE proteome was detected in the SC proteome), but 225 proteins were significantly regulated, either increased or decreased in expres-sion between the two environments (permutation-based false discov-ery rate of 2%, S0=0.1; Fig. 4b and Supplementary Tables 8 and 9). The significantly differentially expressed proteins were enriched for neuron and synapse function (Fig. 4c) and included many impor-tant signaling molecules, especially those related to protein transla-tion (e.g., MTOR, Raptor, Eif4a1 and Eef2) and degradation (e.g., Ubr4, Ubqln1, Psma6 and Trim3) (Fig. 4b and Supplementary Fig. 9a,b). Additionally, proteins involved in the synaptic vesicle cycle (e.g., Cplx2, Syt12, Vamp1, Amph and SNCA), synapse-organizing cell adhesion molecules and scaffold proteins (e.g., Shank1, Ctnnd2, Ncam2, and Dlg2), and GABA receptor subunits (Gabbr1 and 2) were regulated (Fig. 4d). Atl1 and Itpka, for example, have been previously linked with alterations in spine morphology and dynamics23. Taken together, these data demonstrate that genetically targeted protein labeling in MetRS* mice possesses sufficient sensitivity to capture adaptive changes in cell-type-specific proteomes in vivo.

Application of our conditional MetRS* expression enabled us to purify and identify 3,969 proteins in two neuronal types labeled in vivo over a time course of 21 d. Importantly, we show that ANL crosses the blood–brain barrier and that reliable protein labeling can be achieved in animals heterozygous for the MetRS* transgene, which will greatly facilitate the use of the mice in a broad array of Cre-drivers or disease models. A wealth of isotopic metabolic labeling and quan-titative proteomics approaches have recently been developed (e.g., SILAC)24 that can track increasingly subtle changes in proteome dynam-ics and post-translational modifications25. These isotopic labeling tech-niques are compatible with the use of non-canonical amino acids and BONCAT26–28 and can thus now, in principle, be targeted to specific cell types in vivo in R26-MetRS* mice. Alternatively, or in combination, mice could be raised on an ANL diet and then the conditional expression of MetRS* could be driven by a drug-inducible system, or in neurons, via an activity-dependent promoter, enabling a comparison of nascent protein expression in different control and experimental animals.

Using our CaMK2a-Cre::R26 mouse, we identified with cell-type resolution 225 proteins that changed expression levels in excitatory hippocampal neurons following environmental enrichment. Not sur-prisingly, many of the identified proteins are associated with synapses or their regulation. Intriguingly, proteins associated with protein synthesis and degradation were up- or downregulated. Both protein synthesis and degradation are of critical importance in the remod-eling of synaptic proteomes, especially associated with information storage29. A previous study in rodents30 examined environment-induced changes in hippocampal proteins using two-dimensional gels, but unique protein quantification was not confirmed with mass spectrometry; we nevertheless identified several potentially overlap-ping proteins (Supplementary Table 8). It is notable that the proteins we identified reside in excitatory hippocampal neurons, which are crucial for spatial navigation and spatial memory31. The excitatory neurons in hippocampal area CA1 represent an animal’s location in the environment32. The proteomic changes we identify in the circuits responsible for sensing and encoding the spatial environment might suggest that ‘perception’ of the environment is altered as a result of the enriched experience.

We found multiple brain-disease-related proteins in the hippocam-pal excitatory neuron proteome. Surprisingly, a large fraction of the proteins (~30%) that were regulated by exposure to an enriched envi-ronment were also associated with disease. Most disease-related pro-teins are expressed in numerous cell types and tissues, complicating any analysis that endeavors to examine whether disease phenotypes

95 1,3821,382

Cerebellar inhibitory

Hippocampalexcitatory

a

70

log2 fold difference

b

0

1

2

3

4

5

6

7

–4 –3 –2 –1 0 1 2 3 4

Cyfip2

Dlg4

Dmxl2

Map2

Dync1h1Syn1

Itpr1

Calb1

Sept4

Anxa5

Hippocampal excitatory 180 proteins

Cerebellar inhibitory 188 proteins

–log

10 P

-val

ue

c CaMK2a Gad2

6

Norm

alized log2 LF

Q intensity

7

8

9

10

11

12

13

14

15

16

17

d

–15

–10

–5

0

5

10

15

20

25

30

Com

pone

nt 3

(9%

)

–50

–40

–30

–20

–10 0 10 20 30 40 50 60

Component 1 (64.5%)

Cerebellum

Hippocampus

Gad2 cereb

Camk2a hippo

Glia

Figure 3 Cell-type-specific proteomics in vivo. Comparison of hippocampal excitatory neuron and the cerebellar inhibitory neuron proteomes. (a) Venn diagram depicting the numbers of proteins found only in the hippocampal excitatory neuron proteome, the cerebellar inhibitory neuron proteome or both proteomes. (b) Relative abundance of proteins found in both neuronal types in both proteomes (Volcano plot). Shown are fold-differences of protein relative expression (x axis), statistical significance of the enrichment (y axis) and detection threshold based on false-discovery rates (permutation-based) of 2%, S0 = 0.1; thin black lines. Shown in colored dots are proteins significantly enriched in the hippocampal excitatory proteome (left side) or the Purkinje neuron proteome (right side), proteins expressed at similar levels (gray dots). The top five proteins with highest enrichment in each group are labeled. (c) Hierarchical clustering of proteins common to hippocampal excitatory neuron and cerebellar inhibitory neuron proteome showing the net separation (see upper brackets) of the multiple biological replicates (n = 3) analyzed for each neuronal type. LFQ, label-free quantification. (d) Principal component analysis (PCA) showing the net segregation of the hippocampal excitatory neuron and cerebellar inhibitory neuron proteomes, whole hippocampus and cultured glial cells (3–6 biological replicates each) based on two components accounting for ~64% and ~9% of the variability in protein expression levels. See Supplementary Figures 6–8.©

201

7 N

atu

re A

mer

ica,

Inc.

, par

t o

f S

pri

ng

er N

atu

re. A

ll ri

gh

ts r

eser

ved

.

1200 VOLUME 35 NUMBER 12 DECEMBER 2017 nature biotechnology

l e t t e r s

a

Standard Enriched environment

0

1

2

3

4

5

6

7

–3 –2 –1 0 1 2 3

EE 120 proteins EE 105 proteins

Rptor

Tom1l2

Soga3 Gabbr1

Usp7

Psmc3

Vps13c

Rab3b

Ubr4

Acap

Acat 1

Dnpep

Vdac1Prdx2

Vcpip1

Sod1Hnrnpa0

SgtaDctn2

b

Mif

Pccb

log2 fold difference

–log

10 P

-val

ue

c

T

T

T

Endocytosis

Uncoating

Exocytosis

Mitochondria

Endosomefusion

Vesicleacidification

Neurotransmitteruptake

Vamp1

Cytoskeleton

Cplx2

Snph

Brsk1

Sh3gl/2/b2

Atp2b4

Cytoskeleton

Endosomefusion

Dlg2

AmphIgsf8Ctnnd/2/a2

Nos1Ank3

Calb1

Shank1

Dlgap2

Sipa1l1

Itsn1

Pacsin1

Dbn Atcay

Rptor

Ncam2

Fus

Rab3b/c

Sgta

Cacna2d3

Ca2+

Add2

Bsg

Sptbn

Arpc3

d

Gabbr1

Gabbr2

Bin1

Snca

Ptprz1

Syne1

Cntnap2

Itpr1

Ywha/b/e/q/z

Adgrb/2/3

Trim3

Nedd4

Atl1

MTor Rufy3

Upf1

Rapgef4

Iqsec2Rasgrf1Csnk2a1

Ywha/b/e/q/z

Ca2+

Cytosol

Neuron projection

Plasma mbr. bounded cell projection

Vesicle

Cytoplasmic part

Cell projection

Spectrin

Spectrin-associated cytoskeleton

Synapse part

Neuron part

Extracellular region part

Extracellular vesicle

Dendrite

Extracellular organelle

Extracellular exosome

Enrichment–log10 (P-value)

Gene ontology 0 2 4 6 108

Figure 4 The CaMK2a excitatory hippocampal proteome in an enriched environment (EE). (a) Top-down and side-views of the standard cage and EE cages. In the EE cage, a maze (with changeable side walls) was present on the top floor and toys of different sizes, textures and colors were added. (b) Relative abundance of proteins detected in both the standard cage and EE proteome (Volcano plot). Fold-differences of relative protein expression (x axis) is plotted against the statistical significance of the detection (y axis). The significance threshold is based on false-discovery rates (permutation-based) of 2%, S0=0.1; thin black lines. Proteins significantly de-enriched (left side) or enriched (right side) in the EE proteome are shown in dark purple and magenta, respectively (non-differentially expressed proteins in gray). The top ten proteins with highest enrichment in each group are labeled. (c) Gene ontology analysis. Selected cellular processes significantly over-represented in the EE hippocampal excitatory neuron proteome (−log (P values)). Note the presence of multiple terms related to neuronal compartments and synapses (bold). (d) Scheme of presynaptic terminal and postsynaptic compartment showing proteins important for synaptic function or the regulation of synaptic function that were significantly regulated following exposure to the EE (proteins obtained from the Volcano plot in b). See Supplementary Figures 6, 7 and 9.

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

nature biotechnology VOLUME 35 NUMBER 12 DECEMBER 2017 1201

l e t t e r s

arise from particular cell types. Our findings show that within exci-tatory hippocampal neurons, not only were synaptic-disease-related proteins regulated by the enriched environment but also proteins involved in mRNA processing, protein translation and degradation, mitochondrial function and oxidative stress. We conclude that a sur-prisingly high proportion of proteins regulated in excitatory neurons by an enriched environment contribute—when misregulated—to a cohort of different brain diseases (Supplementary Fig. 9b).

With increasing sequencing depth, it has become clear that many genes thought to be specifically expressed in a particular cell type are also present at detectable levels in other cells. In our experiments we focus on the enrichment or de-enrichment of proteins in different samples. To refine the molecular resolution of cellular phenotype, combination of mRNA transcripts with spatial localization can pro-vide deeper phenotyping33,34. The addition of our targetable proteome identification strategy will add a spatial and temporal dimension to our understanding of cellular identity and proteome dynamics in both normal and diseased tissue.

MeThOdSMethods, including statements of data availability and any associated accession codes and references, are available in the online version of the paper.

Note: Any Supplementary Information and Source Data files are available in the online version of the paper.

ACkNowLEDGMENTSWe thank H. Geptin, D. Vogel, N. Fuerst, I. Wüllenweber and F. Rupprecht for excellent technical assistance. We thank E. Noll for the synthesis of ANL and P. Landgraf for the synthesis of the SH-alkyne. We thank S. Garg for her help with FUNCAT and M. Heumueller for his help with some of the experiments. We thank R. Pieaud and S. Junek for their assistance with imaging. We thank F. Kretschmer for his help with the open field analysis. We thank E. Northrup, S. Zeissler, S. Gil Mast and the animal facility of MPI for Brain Research for their excellent support. We thank J. Sanes and J. Chakkalakal for the early generation of a Thy-1 MetRS* mouse. Work in the laboratory of E.M.S. is supported by the Max Planck Society, the European Research Council, DFG CRC 902 and 1080, and the DFG Cluster of Excellence for Macromolecular Complexes. B.A. was supported by a Marie Curie Intra-European Fellowship for career development. C.H. was supported by a Marie Curie Career Integration Grant. D.C.D. is supported by DFG CRC 779 and 854. Research on proteomic labelling at Caltech is supported by the Programmable Molecular Technology Initiative of the Gordon and Betty Moore Foundation.

AUTHoR CoNTRIBUTIoNSB.A.-C., C.T.S. and C.H.: conception and design of experiments, acquisition, analysis and interpretation of data. C.G., S.t.D. and A.R.D.: conception and design of experiments, acquisition of data. I.B., B.N.-A. and E.C.: acquisition of data. A.M.: provided reagents. D.D.: provided reagents and advice on experiments. D.A.T.: conception and design of experiments. J.D.L.: acquisition of data, analysis and interpretation of data. E.M.S.: conception and design of experiments, analysis and interpretation of data, drafting, writing and revising the article. All authors contributed to the writing and revision of the manuscript.

COMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.

reprints and permissions information is available online at http://www.nature.com/reprints/index.html. Publisher’s note: springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

1. Laughlin, S.T., Baskin, J.M., Amacher, S.L. & Bertozzi, C.R. In vivo imaging of membrane-associated glycans in developing zebrafish. Science 320, 664–667 (2008).

2. Dieterich, D.C., Link, A.J., Graumann, J., Tirrell, D.A. & Schuman, E.M. Selective identification of newly synthesized proteins in mammalian cells using bioorthogonal noncanonical amino acid tagging (BONCAT). Proc. Natl. Acad. Sci. USA 103, 9482–9487 (2006).

3. Elsässer, S.J., Ernst, R.J., Walker, O.S. & Chin, J.W. Genetic code expansion in stable cell lines enables encoded chromatin modification. Nat. Methods 13, 158–164 (2016).

4. Mahdavi, A. et al. Engineered aminoacyl-tRNA synthetase for cell-selective analysis of mammalian protein synthesis. J. Am. Chem. Soc. 138, 4278–4281 (2016).

5. Ngo, J.T. et al. Cell-selective metabolic labeling of proteins. Nat. Chem. Biol. 5, 715–717 (2009).

6. Erdmann, I. et al. Cell-selective labelling of proteomes in Drosophila melanogaster. Nat. Commun. 6, 7521 (2015).

7. Yuet, K.P. et al. Cell-specific proteomic analysis in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 112, 2705–2710 (2015).

8. Link, A.J. et al. Discovery of aminoacyl-tRNA synthetase activity through cell-surface display of noncanonical amino acids. Proc. Natl. Acad. Sci. USA 103, 10180–10185 (2006).

9. Miyazaki, J. et al. Expression vector system based on the chicken beta-actin promoter directs efficient production of interleukin-5. Gene 79, 269–277 (1989).

10. Zambrowicz, B.P. et al. Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells. Proc. Natl. Acad. Sci. USA 94, 3789–3794 (1997).

11. Tsien, J.Z. et al. Subregion- and cell type-restricted gene knockout in mouse brain. Cell 87, 1317–1326 (1996).

12. Taniguchi, H. et al. A resource of Cre driver lines for genetic targeting of GABAergic neurons in cerebral cortex. Neuron 71, 995–1013 (2011).

13. Dieterich, D.C. et al. In situ visualization and dynamics of newly synthesized proteins in rat hippocampal neurons. Nat. Neurosci. 13, 897–905 (2010).

14. tom Dieck, S. et al. Direct visualization of newly synthesized target proteins in situ. Nat. Methods 12, 411–414 (2015).

15. Howarth, C., Peppiatt-Wildman, C.M. & Attwell, D. The energy use associated with neural computation in the cerebellum. J. Cereb. Blood Flow Metab. 30, 403–414 (2010).

16. Cho, K.O., Hunt, C.A. & Kennedy, M.B. The rat brain postsynaptic density fraction contains a homolog of the Drosophila discs-large tumor suppressor protein. Neuron 9, 929–942 (1992).

17. Innocenti, M. et al. Abi1 is essential for the formation and activation of a WAVE2 signalling complex. Nat. Cell Biol. 6, 319–327 (2004).

18. Dateki, M. et al. Neurochondrin negatively regulates CaMKII phosphorylation, and nervous system-specific gene disruption results in epileptic seizure. J. Biol. Chem. 280, 20503–20508 (2005).

19. Han, K. et al. Fragile X-like behaviors and abnormal cortical dendritic spines in cytoplasmic FMR1-interacting protein 2-mutant mice. Hum. Mol. Genet. 24, 1813–1823 (2015).

20. Weedon, M.N. et al. Exome sequencing identifies a DYNC1H1 mutation in a large pedigree with dominant axonal Charcot-Marie-Tooth disease. Am. J. Hum. Genet. 89, 308–312 (2011).

21. Novak, M.J. et al. An ITPR1 gene deletion causes spinocerebellar ataxia 15/16: a genetic, clinical and radiological description. Mov. Disord. 25, 2176–2182 (2010).

22. Bennett, E.L., Diamond, M.C., Krech, D. & Rosenzweig, M.R. Chemical and anatomical plasticity brain. Science 146, 610–619 (1964).

23. Shih, Y.T. & Hsueh, Y.P. VCP and ATL1 regulate endoplasmic reticulum and protein synthesis for dendritic spine formation. Nat. Commun. 7, 11020 (2016).

24. Ong, S.E. et al. Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol. Cell. Proteomics 1, 376–386 (2002).

25. Monetti, M., Nagaraj, N., Sharma, K. & Mann, M. Large-scale phosphosite quantification in tissues by a spike-in SILAC method. Nat. Methods 8, 655–658 (2011).

26. Eichelbaum, K., Winter, M., Berriel Diaz, M., Herzig, S. & Krijgsveld, J. Selective enrichment of newly synthesized proteins for quantitative secretome analysis. Nat. Biotechnol. 30, 984–990 (2012).

27. Liu, R. et al. Quantitative non-canonical amino acid tagging (QuaNCAT) proteomics identifies distinct patterns of protein synthesis rapidly induced by hypertrophic agents in cardiomyocytes, revealing new aspects of metabolic remodeling. Mol. Cell. Proteomics 15, 3170–3189 (2016).

28. McShane, E. et al. Kinetic analysis of protein stability reveals age-dependent degradation. Cell 167, 803–815 (2016).

29. Alvarez-Castelao, B. & Schuman, E.M. The regulation of synaptic protein turnover. J. Biol. Chem. 290, 28623–28630 (2015).

30. McNair, K., Broad, J., Riedel, G., Davies, C.H. & Cobb, S.R. Global changes in the hippocampal proteome following exposure to an enriched environment. Neuroscience 145, 413–422 (2007).

31. Moser, E.I., Kropff, E. & Moser, M.B. Place cells, grid cells, and the brain’s spatial representation system. Annu. Rev. Neurosci. 31, 69–89 (2008).

32. O’Keefe, J. & Dostrovsky, J. The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Res. 34, 171–175 (1971).

33. Achim, K. et al. High-throughput spatial mapping of single-cell RNA-seq data to tissue of origin. Nat. Biotechnol. 33, 503–509 (2015).

34. Cembrowski, M.S., Wang, L., Sugino, K., Shields, B.C. & Spruston, N. Hipposeq: a comprehensive RNA-seq database of gene expression in hippocampal principal neurons. eLife 5, e14997 (2016).

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

nature biotechnology doi:10.1038/nbt.4016

ONLINe MeThOdSPlease also see the Life Sciences Reporting Summary for detailed information on experimental design and reagents.ANL synthesis. Azidonorleucine (ANL) was synthesized as previously described4.

Constructs, cell culture and transfection. The mouse MetRS ORF (I.M.A.G.E clone 6416029, GenBank ID BC079643, ATCC) was cloned by PCR into pEGFP-C1 (Clontech). GFP-MetRS L274G (MetRS*) was generated by site-directed mutagenesis (Stratagene). GFP-2A-MetRS* was generated by inserting a sequence encoding for the foot-and-mouth disease virus 2A peptide sequence35 between GFP and MetRS* (Genscript). HEK, COS7 and HeLa cells (ATCC) were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen) sup-plemented with 10% FCS (FCS) at 37 °C in a 5% CO2 atmosphere and transfected with X-tremeGENE9 (Roche) according to the manufacturer’s instructions. Lack of cell line contamination with mycoplasma was checked by PCR (eMyco detec-tion kit, Intron Biotechnology). Experiments were performed 24–48 h post-transfection. For metabolic labeling, cells were incubated in methionine-free DMEM (Invitrogen) supplemented with FCS and 4 mM l-methionine (Sigma), or ANL for 20 min to 5 h, 40 µM anisomycin was added as indicated.

Antibodies. The following antibodies were used for immunofluorescence labe-ling (IF) and/or immunoblotting (IB) at the indicated dilutions: mouse anti-biotin (IF and IB, 1:1,000, Sigma) or rabbit anti-biotin (IF and IB, 1:1,500, Cell Signaling) or rabbit anti-GFP (IF and IB 1:750, Invitrogen) and/or chicken anti-GFP (IF, 1:500; IB: 1:1,000; Aves), rabbit anti-parvalbumin (IF Swant 1:5,000), goat anti-chicken Alexa 647 (IF, 1:750; IB, 1:7,500; Invitrogen), goat anti- rabbit Rhodamine RX (IF, 1:800, Jackson laboratory), goat anti-rabbit FITC (IF, 1:800, Jackson laboratory), goat anti-mouse or anti-rabbit IR680 or IR800 (IB, 1:10,000, Licor), Gp anti-MAP2 (1:1,000, SYSY). Antibodies used for the PLA experiments: rabbit anti-Lamin 1:1,000 (Abcam), rabbit anti-Debrin 1:500 (Sigma), rabbit anti-GluA1 1:500 (Abcam). Mouse anti-far red (1:100, Abcam). In IF experiments DAPI was added to the secondary antibody solutions.

Organotypic slice culture and viral transduction. 450 µm hippocampal slices were prepared from P5-P6 pups with a McIllwain tissue chopper and main-tained, as previously described by others36. Slices were virally transfected after 24 h in vitro with a Cre-expressing recombinant adeno-associated virus (CMV mCherry-2A-Cre rAAV, serotype 1:2, Open BioSystems). For metabolic labe-ling, 4 mM methionine or ANL were added to the standard slice maintenance medium36 for up to 9 d.

Transgenic animals. The cassette - STOPFLOX - GFP-2A-MetRS*- expressed under the control of the CAG actin-derived promoter knock-in in the mouse ROSA26 locus10 was performed by recombination-mediated cassette exchange (RMCE)37 in embryonic stem cells (Taconic GmbH). For activation of the transgene in CA1 pyramidal neurons, homozygote STOPflox R26-MetRS* ani-mals were crossed with homozygote CaMK2a-Cre T29-1 mice11 (Jackson labo-ratory, line #005359) or Gad2-Ires-Cre12 (Jackson Laboratory, line #010082). Genotyping was done by PCR (see Supplementary Table 10 for primer list and size of expected bands). All the procedures involving animals were performed under protocols compliant with and approved by the Hessen Animal Welfare Authority (protocols V54-19c20/15-F122/14 and V54-19c20/15-F126/1012) and the Max Planck Society.

Acute hippocampal slices. 450–500 µm hippocampal slices were prepared from adult animals (6 to 7 weeks old) with a McIllwain tissue chopper and incubated at 30 °C in artificial cerebral spinal fluid (ACSF; in mM, NaCl 125, NaHCO3 25, KCl 2.5, NaH2PO4 1.25, glucose 25, CaCl2 2, MgCl2 1) oxygen-ated by a 95/5 (v/v) O2/CO2 gas mixture. Slices were metabolically labeled for 3–4 h with 4 mM methionine or ANL in the same medium or, when the metabolic labeling was conducted in vivo, directly processed for FUNCAT, as described below.

Fluorescent non-canonical amino-acid tagging (FUNCAT) and immunola-beling. After metabolic labeling in vivo mice were euthanized after anesthe-sia with isofluorane and transcardially perfused with PBS supplemented with

20 mM methionine. Acute or organotypic slices were rinsed twice in ACSF and fixed in 4% PFA in PBS for 20 min at RT. Fixed slices were then washed in PBS and stored at 4 °C in the same buffer until processed further. Slices were then mounted in 4% agar (w/v) and resliced at 30 or 50 µm with a vibratome (Leica), incubated overnight at 4 °C in blocking buffer consisting of 0.5% Triton X100 and 10% goat serum in PBS, 5% sucrose, 2% fish skin gelatin. Samples were washed two times in PBS pH 7.8 and “clicked” overnight with biotin-alkyne (PEG4 caboxamide-propargyl biotin, Molecular Probes) as previously described4. After extensive washing, slices were incubated overnight at 4 °C with anti-GFP and anti-biotin primary antibodies in a twofold dilution of the blocking buffer, washed in 0.1% Triton in PBS and incubated for 3 h at RT with secondary antibodies and DAPI in diluted blocking buffer. Slices were then washed in PBS/Triton and then in PBS and mounted on microscope slides. Fluorescence imaging was performed with a LSM780 laser scanning confocal microscope (Zeiss) with ×20 (NA 0.8) or ×40 (NA 1.4) objectives with appropri-ate excitation laser lines and spectral detection windows.

MetRS* transfection, metabolic labeling and FUNCAT-PLA in cultured neu-rons. Neurons were transfected with Magnetofectamine (Life Technologies) with 1.5 µg of DNA per 40 K neurons for 25 min. One day after transfection the neurons were starved for 5 min in methionine-free Neurobasal A, B27, Glutamax and labeled for 1 h in methionine-free Neurobasal A, B27, Glutamax and 4 mM ANL. After the labeling the neurons were incubated for 10 min in Neurobasal A, B27, Glutamax, washed 2× with PBS, fixed in 4%PFA for 20 min and permeabilized with 0.5% Triton-X100 in 4% of goat serum for 15 min. The FUNCAT reaction was performed overnight at RT. using 0.4 µM of the far-red alkyne (Thermo A10278). After the click reaction, the neurons were blocked for 1 h and incubated for 1.5 h at RT with the corresponding antibody. The FUNCAT PLA protocol was performed as described in ref. 14. In brief, after the FUNCAT protocol, cells were blocked at RT for 1 h (blocking buffer (BB): 4% goat serum in PBS pH7.4), and incubated with the first antibodies during 1.5 h at RT, then washed 3×, 5 min with PBS pH7.4 at RT. PLA oligo-coupled second-ary antibodies (Sigma: DU092004/2) were diluted 1:10 in BB and incubated for 1 h at 37 °C. Probe ligation was performed during 30 min at 37 °C and washed 5× with buffer A (Sigma: 82047). Amplification buffer (Sigma; DU092008) was added for 90 min. Cells were washed quickly twice, then 1 ×10 min with buffer B (Sigma: 82048), and 2× quickly with PBS. Neurons were post-fixed for 10 min (PBS pH7.4, 1 mM MgCl2, 0.1 mM CaCl2, 4% PFA, 4% w/v sucrose) washed 5× with PBS and blocked for 1 h. Alexa-conjugated secondary antibodies were added for 30 min, washed 3 × 5 min with PBS and imaged.

ANL administration to living animals. To achieve a deeper protein labeling for the generation of neuronal-specific proteomes 6-week-old WT or R26-MetR* mice were fed a low methionine diet (diet containing 0.1% methionine) (Ssniff, Germany) beginning 1 week before labeling and were then provided with 60 mM ANL for 21 d. ANL was provided in drinking water supplemented with 7% l-maltose (Sigma) to increase water intake38, resulting in an average daily ANL intake of 0.9 ± 0.1 (mean ± s.d.) mg/day/g body weight.

Open field experiments. After feeding the mice with ANL for 21 d, individual animals were deposited in the center of the open field arena (100 × 100 × 30 cm). The arena was divided in two different equally sized areas (inner and outer). The light intensity was adjusted to 180 lx. Motor behavior movement was evaluated during the 10 min after the initial deposition of the mouse into the arena. Gad2-Cre::MetRS* and CaMK2a-cre::MetRS* were compared with WT mice drinking ANL and with WT mice raised in the same conditions as the ANL-drinking animals.

Enriched environment. Mice (6 weeks of age) were raised in Marlau cages during the 21 d of the ANL-labeling. The labyrinth, running wheels and toys were changed every 2 d to increase novelty. Small pieces of food were hidden into the toys or bedding material to encourage exploratory behavior. Odors were incorporated in toys and small pieces of food using Vanilla flavoring (McCormick).

Electrophysiology. 500 µm hippocampal slices were prepared from 5-week-old MetRS* or WT mice using a Leica vibratome in cold ACSF-sucrose solution

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

nature biotechnologydoi:10.1038/nbt.4016

(concentrations in mM as follows: NaCl 87; NaHCO3 25; NaH2PO4 1.25; KCl 2.5; glucose 10; sucrose 75; CaCl2 0.5; MgCl2 7). Slices were recovered submerged in oxygenated (95/5 O2/CO2) recovery/recording ACSF (con-centrations in mM as follows: NaCl 125; NaHCO3 25; NaH2PO4 1.25; KCl 2.5; Glucose 10; CaCl2 2; MgCl2 1) for 1 h. The slices were transferred to a recording set-up, and recovered for an additional hour in an interface cham-ber perfused with carbogenated ACSF. A stimulating electrode and recording electrode were placed in stratum radiatum. During the recordings, different stimulation intensities were used to obtain the input-output curves.

Pulse-chase turnover experiment. HeLa cells were transfected with CMV-MetRS*-P2A-GFP. Transfection efficiency was evaluated under the micro-scope; dishes in which ~50% of the cells were transfected were used for the experiment. 2 d after transfection the cells were labeled with 80 µCi S35 Cysteine per million of cells ± 8 mM ANL in methionine-free DMEM. Cells were labeled for 1 h or 2 h, to evaluate protein synthesis. After 2 h of labeling media was washed out and replaced by DMEM with 50 µM CHX (Sigma-Aldrich), protein degradation was studied after 6 and 12 h chase periods.

BONCAT, immunoblotting and protein isolation for mass spectrometry. Tissue was homogenized and lysed in PBS supplemented with 1% (w/v) Triton X100 and 0.4% (w/v) SDS, protease inhibitors (PI, 1:750 dilution of protease inhibitor cocktail 3 w/o EDTA, Calbiochem) and benzonase (Sigma, 1:1,000) at 75 °C for 15 min. Lysates were then cleared by centrifugation and stored at −80 °C until used. BONCAT was performed as previously described39. In brief, 90 µg proteins were treated in 120 µL PBS pH 7.8 supplemented with 0.08% (w/v) SDS, 0.2% (w/v) Triton, 300 µM Triazol (Sigma, ref. 678937), 50 µM biotin-alkyne tag (Thermo, ref. B10185) and 83 µg/mL CuBr (prepared by dilution of fresh 10 mg/mL solution in DMSO) at 4 °C overnight in the dark. Biotinylated proteins were then separated by electrophoresis and immunob-lotted with anti-GFP and anti-biotin antibodies (see above) or directly with streptavidin-Alexa 800 (1:10,000 dilution, Invitrogen, ref. Q10173MP). All full-length gels are shown in Supplementary Figure 10.

For mass spectrometry (MS) analyses, tissue was lysed in PBS supplemented with 0.5% Triton, 1% SDS, PI and benzonase, heated at 75 °C for 15 min and cleared by centrifugation. Cleared extracts were alkylated twice with 20 mM iodoacetamide (3 h at RT in the dark) and then cleaned by two passages on PD-SpinTrap G-25 columns (GE healthcare). Samples were then clicked in PBS pH7.8 supplemented with 0.06% SDS, 0.06% of Triton X-100, PI as described above with a cleavable biotin-S-S-alkyne (probe 14b in ref. 40). After click chemistry, excess biotin-alkyne was removed on PD-SpinTrap G-25 columns. Tagged proteins were then affinity-purified with high-capacity Neutravidin agarose beads (Thermo) in PBS supplemented with 1% Triton, 0.15% SDS and PI (binding buffer) overnight at 4 °C. The beads were then washed extensively at RT, first in binding buffer, then PBS + PI, 50 mM ammonium bicarbonate + PI and finally eluted by incubation with 5% β-mercaptoethanol, 0.03% SDS and PI for 30 min at RT. Two consecutive elutions were performed. Eluted proteins were lyophilized, resuspended in water and processed for MS as out-lined below.

MS sample preparation and analyses. Eluted proteins were processed using a modified FASP workflow41 as described previously42. In brief, protein extracts (whole samples for CaMK2a and Gad2) were cleaned and digested sequentially with Lys-C and trypsin on Microcon-10 filters (Merck Millipore, # MRCPRT010 Ultracel YM-10). Digested samples were desalted using ZipTips according to the manufacturer’s instructions, dried in a Speed-Vac and stored at −20 °C until LC-MS/MS analysis. Dried peptides were dissolved in 5% ace-tonitrile with 0.1% formic acid, and subsequently loaded using a nano-HPLC (Dionex U3000 RSLCnano) on reversed phase columns (trapping column: particle size 3 µm, C18, L = 20 mm; analytical column: particle size = 2 µm, C18, L = 50 cm; PepMap, Dionex/Thermo Fisher Scientific). Peptides were eluted in gradients of water (buffer A: water with 5% v/v dimethylsulfoxide and 0.1% formic acid) and acetonitrile (buffer B: 5% dimethylsulfoxide, 15% water and 80% acetonitrile (v/v/v) and 0.08% formic acid). All LC-MS-grade solvents were purchased from Fluka. Gradients were ramped from 4% to 48% B in 178 min at a flowrate of 300 nl/min. Peptides eluting from the column were ionized online using a Thermo nanoFlex ESI-source and analyzed in a Thermo

“Q Exactive Plus” mass spectrometer or a Thermo “Orbitrap Elite” mass spec-trometer. Mass spectra were acquired over the mass range 350–1,400 m/z (Q Exactive Plus) or 350-1,600 m/z (Orbitrap Elite) and sequence information was acquired by computer-controlled, data-dependent automated switching to MS/MS mode using collision energies based on mass and charge state of the candidate ions (full parameter set in Supplementary Table 11). All samples were measured in (at least) triplicate LC-MS runs and the biological replicates (one mouse represents one replicate) for each group are noted in the respec-tive figure legends.

Raw MS data were processed and analyzed using MaxQuant43 (see param-eters in Supplementary Table 11). In brief, spectra were matched to a Mus musculus database downloaded from uniprot.org (reviewed and non-reviewed, downloaded on May 13, 2016) and a contaminant and decoy database. Precursor mass tolerance was set to 4.5 p.p.m., fragment ion tolerance to 20 p.p.m. (QExactive) or 0.5 Da (Orbitrap Elite), with fixed modification of Cys residues (carboxyamidomethylation +57.021 Da) and variable modifications of Met residues (Ox +15.995 Da), Lys residues (acetylation +42.011 Da), Asn and Gln residues (deamidation +0.984 Da) and of N termini (carbamylation +43.006 Da). Peptide identifications were calculated with FDR < 0.01, and proteins with one peptide per protein included for subsequent analyses. All proteomic data associated with this study have been deposited to the ProteomeXchange Consortium via the PRIDE44 partner repository with the data set identifier PXD007703. Peptide intensities were analyzed with MaxQuant and Perseus45 (normalization of the intensities: mean centering; Volcano plots: two-sided t-test, unpaired; segmentation analysis: number of clusters: 300, maximal number of iterations: 10). Proteins only identified by reverse peptides were removed. Protein intensities and LFQ label-free quantification intensities were averaged over technical triplicates (representing different runs of the same biological replicate) and log 2 transformed. The full proteomes (cerebellum, hippocampus, glial cells) were analyzed by using proteins only quantified by ≥2 biological replicates. The protein intensities in each proteome were normalized by adjusting the mean protein intensity (measured over all biological replicates) to the value “10”. Proteins purified in the MetRS samples of the CaMK2a and Gad2 replicates were compared to the corresponding WT samples. The fold enrichment of each protein was calculated by subtracting the non-normalized MetRS intensity from the intensity measured in the paired WT sample. Only proteins detected exclusively in the MetRS samples or found to be ≥3 times enriched in MetRS in three replicates or more (≥2 replicates in Gad2) were retained for further analysis. The other proteins enriched or exclusively quanti-fied in fewer replicates were analyzed separately to detect “unique” proteins for the distinct proteomes (e.g., CaMK2a against Gad2). Proteins with no enrich-ment but quantified in ≥3 replicates of MetRS and WT were counted as overlap-ping proteins (overlap in Venn diagrams in Figs. 2 and 3 and Supplementary Fig. 8). Proteins enriched or exclusively detected both in different MetRS and WT replicates were also assigned to the overlap. Proteins enriched in MetRS were filtered, retaining only proteins with LFQ intensities in ≥3 replicates (or ≥2 replicates in Gad2). LFQ intensities were normalized by adjusting the mean LFQ intensity of each Gad2 experiment or the mean LFQ intensity of all CaMK2 replicates to “10”. Normalized LFQ intensities of the purified MetRS samples were analyzed against the full proteomes using proteins only quantified in all samples. For principal component analysis (without category enrichment), missing protein intensities were replaced by their respective mean LFQ inten-sity. LFQ intensities of each CaMK2 and Gad2 replicate were mean-centered and adjusted to “10”. Differentially regulated proteins in CaMK2-EE against CaMK2-HC, and Gad2 against CaMK-HC were analyzed by a two-sided t-test (number of randomizations: 250, permutation-based FDR = 2.0%, S0 = 0.1). Common proteins were hierarchically clustered using the Euclidean distance of averaged LFQ intensities (linkage) preprocessed with k-means (number of clusters: 50, max. 10 iterations).

The criteria used for “unique” proteins in the comparative study between Camk2a and the Gad2 proteomes was: for the “unique” proteins in Gad2; proteins that were detected in ≤2 out of 6 replicates in Camk2a proteome. For the “unique” proteins in Camk2; proteins detected in ≤1 out of 3 replicates in Gad2 proteome (Supplementary Table 6). For the Venn diagram in Figure 3, we used a more stringent criteria for the unique protein numbers: for the “unique” proteins in Gad2; proteins that were detected in ≤0 out of 6 rep-licates in Camk2a proteome. For the “unique” proteins in Camk2; proteins

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

nature biotechnology doi:10.1038/nbt.4016

detected in ≤0 out of 3 replicates in Gad2 proteome (Supplementary Table 6). For the comparative study of Camk2-SC and Camk2-EE we found 16 and 12 “unique” proteins, respectively. These proteins were identified in all the replicates of each proteome and completely absent in the other (0 out of 6 replicates). Nevertheless, we did not analyze these proteins further because the corresponding peptides exhibited variable and relatively low intensities.

For the analysis of the methionine content of the purified proteomes (Supplementary Fig. 7c) the number of methionine residues per 100 amino acids was counted from the full protein sequences. The upper and lower bounds of the box indicate the 75 and 25 percentiles of the data, respectively, and the whiskers indicate the 95th (upper) and 5th (lower) percentiles.

Primary glial culture and preparation for MS. The cortices of P1-day-old rat pups were dissected and mechanically dissociated after digestion with papain. Cells were plated on uncoated 6 cm plastic culture dishes (equivalent to 1/5 cortex per dish) and grown at 37 °C in a 5% CO2 atmosphere in minimal essential medium supplemented with glutamax and 10% FCS, and starting after 7 d in vitro (DIV), in Neurobasal-A supplemented with B27 and Glutamax (Life Technologies) pre-conditioned on primary cortical neuronal cultures46. At DIV 18 the cells were washed and harvested in ice-cold PBS supplemented with protease inhibitors. Cells were lysed in 200 mM Tris (pH 8.4), 1M NaCl supplemented with 8 M urea, 4% CHAPS, and Benzonase, cleared by centrifu-gation and finally alkylated and enzymatically digested essentially as described previously40 and stored at −20 °C until LC-MS/MS analysis.

Gene ontology, tracking of cell-type-specific marker distributions and genes related to disease. Gene ontology analyses were performed with Gorilla47 by comparing specific protein subsets (e.g., the CaMK2a hippocampal proteome) to our mass spectrometry data from the tissue of origin (e.g., the hippocampus, Supplementary Table 12) as background, the 30 GO terms most enriched (P value <1 × 10−3) were ranked by ascending statistical significance P values (see Figs. 2e and 4c, Supplementary Fig. 8d and Supplementary Tables 2, 5 and 9). The cell-type-specific markers used for evaluation of our neuronal proteomes (Fig. 2g and Supplementary Fig. 8e) were obtained as follows: a set of 290 glial and 644 neuronal protein markers was obtained by extracting proteins showing a twofold higher abundance in astrocytes, oligodendrocytes and microglia cultures compared with neuronal cultures in48. Unique hippoc-ampal-cell-type markers were taken from reference49. Cerebellar layer-specific markers (i.e., markers enriched in or excluded from the Purkinje neuron layer) were extracted by selecting markers with an enrichment score ≥2 (enriched) or ≤2 (de-enriched) in ref. 50 (see full lists in Supplementary Table 3).

The statistical significance of marker enrichment was assessed based on their relative hypergeometric probability distribution (Supplementary Table 3). The proteins related to diseases that were regulated by EE were collected from VarElect; only terms with scores > 50 were selected. In the case of bipolar dis-orders, genes with scores > 90 were selected due to the large number of genes related to this disease. Some additional terms were manually added based on the literature (Supplementary Fig. 9b).

35. de Felipe, P. et al. E unum pluribus: multiple proteins from a self-processing polyprotein. Trends Biotechnol. 24, 68–75 (2006).

36. Stoppini, L., Buchs, P.A. & Muller, D. A simple method for organotypic cultures of nervous tissue. J. Neurosci. Methods 37, 173–182 (1991).

37. Branda, C.S. & Dymecki, S.M. Talking about a revolution: The impact of site-specific recombinases on genetic analyses in mice. Dev. Cell 6, 7–28 (2004).

38. Bachmanov, A.A., Tordoff, M.G. & Beauchamp, G.K. Sweetener preference of C57BL/6ByJ and 129P3/J mice. Chem. Senses 26, 905–913 (2001).

39. Dieterich, D.C. et al. Labeling, detection and identification of newly synthesized proteomes with bioorthogonal non-canonical amino-acid tagging. Nat. Protoc. 2, 532–540 (2007).

40. Szychowski, J. et al. Cleavable biotin probes for labeling of biomolecules via azide-alkyne cycloaddition. J. Am. Chem. Soc. 132, 18351–18360 (2010).

41. Wisniewski, J.R., Zougman, A., Nagaraj, N. & Mann, M. Universal sample preparation method for proteome analysis. Nat. Methods 6, 359–362 (2009).

42. Schanzenbächer, C.T., Sambandan, S., Langer, J.D. & Schuman, E.M. Nascent proteome remodeling following homeostatic scaling at hippocampal synapses. Neuron 92, 358–371 (2016).

43. Cox, J. & Mann, M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. 26, 1367–1372 (2008).

44. Vizcaíno, J.A. et al. ProteomeXchange provides globally coordinated proteomics data submission and dissemination. Nat. Biotechnol. 32, 223–226 (2014).

45. Tyanova, S. et al. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat. Methods 13, 731–740 (2016).

46. Sutton, M.A. et al. Miniature neurotransmission stabilizes synaptic function via tonic suppression of local dendritic protein synthesis. Cell 125, 785–799 (2006).

47. Eden, E., Navon, R., Steinfeld, I., Lipson, D. & Yakhini, Z. GOrilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinformatics 10, 48 (2009).

48. Sharma, K. et al. Cell type- and brain region-resolved mouse brain proteome. Nat. Neurosci. 18, 1819–1831 (2015).

49. Zeisel, A. et al. Brain structure. Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq. Science 347, 1138–1142 (2015).

50. Kirsch, L., Liscovitch, N. & Chechik, G. Localizing genes to cerebellar layers by classifying ISH images. PLoS Comput. Biol. 8, e1002790 (2012).

© 2

017

Nat

ure

Am

eric

a, In

c., p

art

of

Sp

rin

ger

Nat

ure

. All

rig

hts

res

erve

d.

1

nature research | life sciences reporting summ

aryJune 2017

Corresponding author(s):

Initial submission Revised version Final submission

Life Sciences Reporting SummaryNature Research wishes to improve the reproducibility of the work that we publish. This form is intended for publication with all accepted life science papers and provides structure for consistency and transparency in reporting. Every life science submission will use this form; some list items might not apply to an individual manuscript, but all fields must be completed for clarity.

For further information on the points included in this form, see Reporting Life Sciences Research. For further information on Nature Research policies, including our data availability policy, see Authors & Referees and the Editorial Policy Checklist.

Experimental design1. Sample size

Describe how sample size was determined. The experiments described in this exploratory study were done for the first time. No pre-specified effect size could be determined a priori. For MS, in general, two replicates are acceptable if the overlap between them is good (e.g. r2 greater than or equal to 0.80). In this study we used a minimum of two replicates per proteome and up to 6 replicates. We had excellent reproducibility between replicates, with r2 greater than 0.90.

2. Data exclusions

Describe any data exclusions. No data were excluded from the study.

3. Replication

Describe whether the experimental findings were reliably reproduced.

The isolation of specific proteomes was reproducible as is shown in the paper and described above. Nevertheless due to the several steps of the purification protocol, the purification of some samples failed and there was no mass spec analysis.

4. Randomization

Describe how samples/organisms/participants were allocated into experimental groups.

As a general rule, mice (wt or transgenic) were all exposed to the same treatment. In the case of EE and HC experiments mice were randomly distributed in each of the cage types.

5. Blinding

Describe whether the investigators were blinded to group allocation during data collection and/or analysis.

Investigators were not blinded.

Note: all studies involving animals and/or human research participants must disclose whether blinding and randomization were used.

Nature Biotechnology: doi:10.1038/nbt.4016

2

nature research | life sciences reporting summ

aryJune 2017

6. Statistical parameters For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or in the Methods section if additional space is needed).

n/a Confirmed

The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.)

A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same sample was measured repeatedly

A statement indicating how many times each experiment was replicated

The statistical test(s) used and whether they are one- or two-sided (note: only common tests should be described solely by name; more complex techniques should be described in the Methods section)

A description of any assumptions or corrections, such as an adjustment for multiple comparisons

The test results (e.g. P values) given as exact values whenever possible and with confidence intervals noted

A clear description of statistics including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range)

Clearly defined error bars

See the web collection on statistics for biologists for further resources and guidance.

SoftwarePolicy information about availability of computer code

7. Software

Describe the software used to analyze the data in this study.

ImageJ64. Prism 6. Origin 2015G. MaxQuant 1.5.3.8. Perseus 1.5.5.3.

For manuscripts utilizing custom algorithms or software that are central to the paper but not yet described in the published literature, software must be made available to editors and reviewers upon request. We strongly encourage code deposition in a community repository (e.g. GitHub). Nature Methods guidance for providing algorithms and software for publication provides further information on this topic.

Materials and reagentsPolicy information about availability of materials

8. Materials availability

Indicate whether there are restrictions on availability of unique materials or if these materials are only available for distribution by a for-profit company.

No unique materials are used.

9. Antibodies

Describe the antibodies used and how they were validated for use in the system under study (i.e. assay and species).

Mouse anti-biotin (IF and IB, 1:1000, Sigma, cat:033700). Rabbit anti-biotin (IF and IB, 1:1500, Cell Signaling, cat:5567). Rabbit anti-GFP (IF and IB 1:750, Invitrogen, cat:A11122). Chicken anti-GFP (IF, 1:500; IB: 1:1000; Aves,cat:1020). Gp anti-MAP2 (1:1000, SYSY,cat:188004). Antibodies used for the PLA experiments: rabbit anti-Lamin 1:1000 (abcam, cat:ab16048), rabbit anti-Debrin 1:500 (sigma, cat:d3816), rabbit anti-GluA1 1:500 (abcam, cat:ab31232). Mouse anti-far red (1:100, abcam, cat:ab52060 ). Secondary antibodies Goat anti-chicken Alexa 647 (IF, 1:750; IB, 1:7500; Invitrogen, a21449). Goat anti-rabbit Rhodamine RRX (IF, 1:800, Jackson laboratory, cat:711295152). Goat anti-rabbit FITC (IF, 1:800, Jackson laboratory, cat:111095003). Goat anti-mouse or anti-rabbit IR680 or IR800 (IB, 1:10.000, Licor,cat:296-32213 and 296-32211). Antibody specificity was evaluated using the proper negative controls.

Nature Biotechnology: doi:10.1038/nbt.4016

3

nature research | life sciences reporting summ

aryJune 2017

10. Eukaryotic cell linesa. State the source of each eukaryotic cell line used. Hek, HeLa and Cos7 were obtained from ATCC.

b. Describe the method of cell line authentication used. None of the lines have been authenticated.

c. Report whether the cell lines were tested for mycoplasma contamination.

Lack of cell line contamination with mycoplasma was checked by PCR (eMyco detection kit, Intron Biotechnology).

d. If any of the cell lines used are listed in the database of commonly misidentified cell lines maintained by ICLAC, provide a scientific rationale for their use.

No commonly misidentified cell lines were used.

Animals and human research participantsPolicy information about studies involving animals; when reporting animal research, follow the ARRIVE guidelines

11. Description of research animalsProvide details on animals and/or animal-derived materials used in the study.

All the mice used in this study were C57BL/6 from Jackson Laboratory. Animals used were 6 weeks old at the beginning of the experiments. Both males and Females were used.

Policy information about studies involving human research participants

12. Description of human research participantsDescribe the covariate-relevant population characteristics of the human research participants.

Study did not involved humans.

Nature Biotechnology: doi:10.1038/nbt.4016