a zebrafish functional genomics model to …...2020/02/23 · 88 many of the mammalian components...

1

A zebrafish functional genomics model to investigate the role of human A20 variants in 2

vivo 3

4

Authors: 5

Cultrone, Daniele1,2; Zammit, W. Nathan1,2; Self, Eleanor1,2; Postert, Benno2,3; Han, Jeremy 6

ZR2,4; Bailey, Jacqueline1,2; Warren, Joanna1,2; Croucher, David R2,4; Kikuchi, Kazu2,5; 7

Bogdanovic, Ozren2,6; Chtanova, Tatyana1,2; Hesselson, Daniel2,3; Grey, Shane T. 1,2† 8

9

Affiliations: 10 1Immunology Division; 3Diabetes Division; 4The Kinghorn Cancer Centre; 6Epigenetics 11

Division; Garvan Institute of Medical Research, 384 Victoria St, Darlinghurst, New South Wales, 12

2010 Australia 13 2St Vincent’s Clinical School, The University of New South Wales Sydney, New South Wales, 14

2010 Australia. 15 5Developmental and Stem Cell Biology Division, Victor Chang Cardiac Research Institute, 16

Darlinghurst, NSW 2010, Australia 17

18

†Corresponding Author and Lead Contact: Email [email protected] (S.T.G.) 19

20

Running title: A new zebrafish model to investigate A20 gene variants in vivo 21

22

23

.CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under apreprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted February 25, 2020. ; https://doi.org/10.1101/2020.02.23.961763doi: bioRxiv preprint

https://doi.org/10.1101/2020.02.23.961763

http://creativecommons.org/licenses/by-nc-nd/4.0/

SUMMARY 24

Germline loss-of-function variation in TNFAIP3, encoding A20, has been implicated in a wide 25

variety of autoinflammatory and autoimmune conditions, with acquired somatic missense 26

mutations linked to cancer progression. Furthermore, human sequence data reveals that the A20 27

locus contains ~400 non-synonymous coding variants which are largely uncharacterised. The 28

growing number of A20 coding variants with unknown function, but potential clinical impact, 29

poses a challenge to traditional mouse-based approaches. Here we report the development of a 30

novel functional genomics approach that utilises the new A20-deficient zebrafish (Danio rerio) 31

model to investigate the impact of TNFAIP3 genetic variants in vivo. Similar to A20-deficient 32

mice, A20-deficient zebrafish are hyper-responsive to inflammatory triggers and exhibit 33

spontaneous early lethality. While ectopic addition of human A20 rescued A20-null zebrafish from 34

lethality, missense mutations at two conserved A20 residues, S381A and C243Y reversed this 35

protective effect. Ser381 represents a phosphorylation site important for enhancing A20 activity 36

that is abrogated by its mutation to alanine, or by a C243Y mutation that associates with human 37

autoimmune disease. These data reveal an evolutionarily conserved role for A20, but also 38

demonstrate how a zebrafish functional genomics pipeline can be utilized to investigate the in vivo 39

significance of medically relevant TNFAIP3 gene variants. This approach could be utilised to 40

investigate genetic variation for other conserved genes. 41

42



https://doi.org/10.1101/2020.02.23.961763


INTRODUCTION 43

In mammals, the cytoplasmic ubiquitin-editing enzyme A20, encoded by the TNFAIP3 gene, plays 44

a key role in maintaining inflammatory homeostasis 1 mediated by its function to inhibit NF-κB 45

activation 2-6. The medical importance of A20’s role in dampening NF-κB is highlighted by cases 46

of germline A20 loss-of-function mutations in humans who present with severe autoinflammatory 47

disease 7,8, and by the impact of A20 deletion in mice which results in spontaneous and widespread 48

NF-κB activation, multi-organ inflammation and premature lethality 1,9. The emergence of A20 49

mutations and coding variants as causal genetic factors driving human disease 10,11 highlight the 50

need to better understand A20’s functional domains controlling inflammation in vivo. 51

52

A20 suppresses NF-κB signalling with inhibitory activity against key molecular substrates 53

including signalling molecules TNF receptor-associated factor 6 (TRAF6) 9, receptor interacting 54

protein 1 (RIP1) 12,13, and the IκB kinase complex (IKK) 14. The A20 ovarian-tumour (OTU) 55

domain exhibits deubiquitinating editing (DUB) activity cantered on Cys103, which cleaves 56

activating K63-linked ubiquitin chains from RIPK1, TRAF6 and NEMO to terminate NF-κB 57

signalling 9,12,13,15. The A20 zinc finger 4 (ZnF4) exhibits E3 ligase activity, adding K48-linked 58

ubiquitin chains to RIPK1, triggering RIPK1 proteolysis 12,16. In addition, the C-terminal ZnF7 59

domain of A20 binds linear ubiquitin to non-catalytically suppress NF-κB activity 17. A20 is 60

regulated at the level of gene transcription 6, whereby NF-κB activation induces TNFAIP3 61

expression 18,19 forming a negative feedback loop, and the inhibitory activities of A20 enzymatic 62

sites are enhanced by IKKB-mediated serine phosphorylation at Ser381 20,21. 63

64

We previously showed that coding variants that impair A20 Ser381 phosphorylation 13, and 65

thereby reduce the activity of both enzymatic Cys103 and ZnF4 domains of A20 12,20, exhibited a 66

much larger in vivo effect in mice than mutations that disable the individual enzymatic sites 67

Cys103 and ZnF4 13,21,22. These data point towards an important role for non-catalytic Ser381 in 68

regulating A20’s anti-inflammatory activity in vivo, however, the in vivo impact of deleting Ser381 69

has not been tested. 70

71

The importance of A20 in human disease, as well as the emerging functional complexities 72

highlighted by experimental studies examining A20 mutations, coupled with the discovery of 73

100’s of new TNFAIP3 coding variants from human genome sequencing studies (e.g. gnomAD) 74

highlight the need for novel approaches to investigate A20 functional domains. We hypothesised 75



https://doi.org/10.1101/2020.02.23.961763


that elucidation of A20’s conserved motifs across additional species may aid the resolution of 76

A20’s functional domains as well as highlight protein domains of relevance to human disease. 77

Here we report the development of a novel functional genomics model that utilises the new A20-78

deficient zebrafish (Danio rerio) model to investigate the impact of TNFAIP3 genetic variants in 79

vivo. Using this zebrafish paradigm we highlight the strong evolutionary conservation of A20’s 80

role to maintain inflammatory homeostasis, and demonstrate how a zebrafish functional genomics 81

approach can be utilized to investigate the in vivo significance of human TNFAIP3 gene variants. 82

This same approach could be utilised to increase understanding of the impact of human genetic 83

variation for other highly conserved genes. 84

85



https://doi.org/10.1101/2020.02.23.961763


RESULTS 86

Genomic analysis of the zebrafish tnfaip3 locus. 87

Many of the mammalian components of the NF-κB signalling system 23,24 and innate immunity 25 88

are conserved in zebrafish making this animal model a candidate in vivo paradigm for the analysis 89

of A20’s key functional motifs. This concept was further supported by analysis of zebrafish 90

genomic data which revealed gene order preservation of the TNFAIP3 locus with OLIG3 and 91

PERP in zebrafish (Figure S1A), and a conserved and easily identifiable Topologically Associating 92

Domain (TADs) overlapping with A20 and highly enriched in the regulatory histone mark 93

H3K27ac (histone 3, lysine 27 acetylation) (Figure S1B), a hallmark of active promoters and 94

enhancers 26,27. Furthermore, some of these conserved H3K27c marked regions overlapped with 95

the positions of A20 SNPs identified in human genome wide association studies (GWAS) to be 96

linked with Crohn’s disease and Systemic lupus erythematosus (Figure S1A). The tnfaip3 97

zebrafish orthologue encoding a predicted 762 amino acid polypeptide comprising an N-terminal 98

ovarian tumour (OTU) domain (identity score of 94/100), including the putative catalytic cysteine 99

at position 103 in A20’s OTU domain 28, and seven zinc finger motifs in the C-terminus (identity 100

scores of 79/100) (Figure S2). The zebrafish A20 C-terminal domain also contained zinc finger 4 101

(ZnF4) and ZnF7 regions, which showed multiple sequence alignment scores of 91/100 and 102

100/100 respectively, and are thus highly homologous to functional ZnF4 and ZnF7 domains 103

identified in mammalian studies 17,21. In mammalian cells A20 expression is controlled at the level 104

of transcription by the NF-κB pathway in response to microbial stimulation of TLR 19, a pathway 105

that represents a key component of innate immunity conserved in zebrafish 29. Stimulation of 106

cultured zebrafish embryos 3 days post-fertilization (dpf) with 75 µg/ml of the gram-negative 107

bacteria product lipopolysaccharide (LPS) also resulted in the rapid induction of zebrafish (zf) 108

zfA20 (tnfaip3) mRNA as well as TNF (tnf), zfIL-1beta (il1b) and zfIL-6 (il6) mRNA (Figure 1 109

A and B). 110

111

Generation of a zebrafish model of A20 deficiency. 112

An optimal zebrafish functional genomics model to test human A20 variants would have an A20 113

null genetic background. Because A20 deletion in zebrafish had not been previously reported, we 114

utilised site directed transcription activator-like effector nucleases (TALEN) to introduce a 115

disabling mutation within exon 2 of the zebrafish A20 locus (Figure 1C). A zebrafish line was 116

identified that showed a 20 bp deletion in the tnfaip3 sequence which was predicted to introduce 117

a premature truncation at amino acid 127 (tnfaip3gi3, hereafter A20∆127; Figure 1C). Heterozygous 118



https://doi.org/10.1101/2020.02.23.961763


A20∆127 zebrafish were bred to homozygosity and deletion of A20 was confirmed in offspring by 119

high-resolution melting assay (HRMA), Sanger sequencing, and analysis of PCR products that 120

differentiate A20 wild-type alleles from A20+/∆127, and A20∆127/∆127 (Figure S3). Heterozygous 121

A20∆127 zebrafish were mated to generate A20+/+, A20+/∆127, and A20∆127/∆127 zebrafish. 122

123

A20 deficiency is lethal in zebrafish. 124

To validate the use of A20 deficient zebrafish for testing human A20 variants, we investigated the 125

impact of zebrafish A20 deletion with respect to sentinel A20 phenotypes reported in mammals. 126

The key mouse phenotypes reported in the literature include loss of organ homeostasis, gross 127

anatomical changes (i.e. runting) and premature lethality 1, as well as hyper-responsiveness to 128

TLR-induced NF-κB activity and heightened macrophage activity 9,30. Accordingly, A20+/+, 129

A20+/∆127, and A20∆127/∆127 zebrafish were followed through time to assess survival. A20∆127/∆127 130

zebrafish presented in the expected Mendelian ratios at 1 week post-fertilization (wpf), however, 131

by 2wpf zebrafish homozygous for A20 deletion presented with premature lethality and no 132

A20∆127/∆127 zebrafish survived to three weeks post-fertilization (Figure 1D). Many A20∆127/∆127 133

zebrafish also exhibited gross abnormalities including reduced body size (Figure 1E) and evident 134

malformations of the torso (Figure 1F). In contrast, but like mice heterozygous for A20 deletion 1, 135

A20+/∆127 zebrafish were viable and showed no gross phenotype or lethality. Mice lacking A20 136

present with a rapid and early onset of systemic inflammation of the organs particularly the liver 137 1. To examine the impact of A20 deletion upon organ homeostasis A20 deficient zebrafish were 138

crossed with 2-Color Liver Insulin acinar Pancreas (2CLIP) zebrafish, which express dsRed 139

fluorescent protein driven by the fatty acid binding protein 10 (fabp10) promoter in hepatocytes 140 31. A20∆127/∆127:2CLIP zebrafish are phenotypically runted but also show a marked reduction in 141

liver area disproportionate to their body size when compared to WT zebrafish (Figure 1G). 142

Reduction in liver size was not due to a developmental defect as liver size was normal in 1 wpf 143

A20∆127/∆127:2CLIP zebrafish. Confocal analysis of A20∆127/∆127:2CLIP zebrafish livers (Figure 144

1H), revealed a reduction in liver cellularity, and an increase in the frequency of caspase-3 positive 145

hepatocytes. 146

147

A20 limits NF-κB activation in zebrafish. 148

To investigate NF-κB activation in A20 deficient zebrafish we crossed A20+/∆127 zebrafish to a 149

NF-κB:EGFP transgenic zebrafish reporter line. The responsiveness of NF-κB activation in 150



https://doi.org/10.1101/2020.02.23.961763


zebrafish in vivo is a direct result of colonisation by environmental microbiota 24. As shown in 151

Figure 2A and B, on a wild type background the NF-κB:EGFP reporter demonstrates spontaneous 152

activation. NF-κB driven fluorescence was found to be at its highest expression in those tissues 153

that mostly interact with the external environment, namely the mouth, neuromasts, pharyngeal 154

tooth and gills (Figure 2A a, b, c, d). Also, NF-κB driven fluorescence increased through time 155

(compare 3dpf A20+/+ zebrafish with 6dpf; Figure 2B) most likely reflecting exposure to 156

environmental microbiota 24, and this was increased further in A20+/∆127 and A20∆127/∆127 157

zebrafish in a gene dose manner (Figure 2B, C and D). Furthermore, the addition of LPS to 158

zebrafish embryos further enhanced NF-κB activation and this was most pronounced in 159

A20∆127/∆127 zebrafish (Figure 2C and D). NF-κB controls the expression of inflammatory 160

response genes, and A20∆127/∆127 zebrafish showed increased basal expression of MPEG1 (Figure 161

2E), a macrophage expressed NF-κB regulated gene 32-34. 162

163

Zebrafish A20-deficient macrophages are hyper responsive to inflammatory triggers 164

A20 loss of function mutations result in macrophage hyper responsiveness to microbial stimulation 165

and spontaneous inflammasome activation 9,13,17. Tissue macrophages play a key role in sensing 166

pathogenic microbiota in zebrafish 33,35. To investigate the impact of A20 deletion on macrophage 167

homeostasis, heterozygous A20+/∆127 zebrafish were crossed to a macrophage reporter line 168

whereby RFP is expressed under the control of the promotor of macrophage-expressed gene 1.1 169

(mpeg1.1) 36. Loss of A20 completely altered both the number and phenotype of tissue 170

macrophages in ways consistent with increased cellular response to stimulation. The number of 171

macrophages detected in the skin was 2-3 fold higher in A20∆127/∆127 zebrafish when compared 172

to their wild type counterparts (Figure 3A and C). Furthermore, the majority of the dermal 173

macrophages in A20∆127/∆127 zebrafish exhibited dendritic extensions (Figure 3A) indicative of an 174

activated state 37,38. This was made further apparent by histological analysis through transverse 175

sections of the cranium, which revealed dense infiltration of mpeg1.1 positive cells close to the 176

dermal surface (Figure 3E). The position of the mpeg1.1 positive cells corresponded with cells at 177

the dermal surface exhibiting bright NF-κB:EGFP reporter activity (Figure 3E). 178

179

Recruitment to sites of inflammation constitute important physiological roles for the macrophage 180

immune compartment 35. Using a tail clip wound model 39 we found that both A20+/∆127 and 181

A20∆127/∆127 zebrafish showed an increased recruitment of macrophages at the wound interface 182



https://doi.org/10.1101/2020.02.23.961763


through time (Figure 3F). Interestingly, this analysis revealed a macrophage phenotype for 183

heterozygous A20+/∆127 zebrafish post challenge whereas we did not observe a change in 184

A20+/∆127 macrophage numbers in the steady state (Figure 3A and C). To investigate this further, 185

we used two-photon imaging of single mpeg1.1:RFP reporter A20+/∆127 zebrafish to track 186

individual macrophage cell movement in vivo in real time (Figure 3G). This analysis revealed that 187

A20+/∆127 macrophages exhibited increased cell movement (Figure 3G), attributed to an increase 188

in their average track speed, displacement and length (Figure 3H). In addition to macrophages, T 189

cell numbers were also increased in A20 deficient zebrafish (Figure 3B and D), a phenotype 190

comparable to A20-deficient mice which exhibit increased numbers of CD3+ lymphocytes 1. 191

192

Human A20 rescues A20∆127/∆127 zebrafish from lethality 193

The phenotypic similarities between A20-deficient zebrafish and those reported for A20-deficient 194

mice 1,9,17, as well as the profound inflammatory syndromes reported for human subjects with 195

heterozygous A20 deficiency 8, supported utilising zebrafish as a model organism to investigate 196

the role of A20 protein domains in regulating its function in vivo. For our functional analysis 197

pipeline, we determined that rescue of A20∆127/∆127 zebrafish from lethality would provide the 198

most robust readout for loss or gain of function phenotypes in various A20 genetic variants. For 199

the functional analysis, zebrafish eggs derived from A20+/∆127 x A20+/∆127 crosses were 200

ectopically injected with wild-type human (h)A20 expressed in the construct Ubb-hA20-P2A-201

EGFP that also expresses GFP. The low integration Ubb-hA20-P2A-EGFP construct was 202

deliberately chosen to limit potential off-target effects as well as false positives 40-42. The zebrafish 203

eggs were injected in a blinded fashion with the Ubb-hA20-P2A-EGFP construct and at 3wpf 204

surviving zebrafish were assessed for their genotype and the presence of GFP fluorescence (Figure 205

S5A). At 1 wpf A20∆127/∆127 zebrafish presented in the expected mendelian ratios of ~25% but 206

exhibited a highly penetrant lethal phenotype, with no survivors at 3wpf (Figure 1D). In contrast, 207

with ectopic expression of hA20, A20∆127/∆127 zebrafish comprised 9.91% of the total genotypes 208

at 3wpf (Figure 4A). Given an expected Mendelian ratio of 25%, ectopic expression of hA20 209

rescued~40% of A20∆127/∆127 zebrafish from lethality. 210

211

Alanine substitution of Ser381 and the human mutation C243Y impair A20’s in vivo function 212

We previously showed that a series of three different A20 coding variants, all within A20’s OTU 213

domain, showed a strong correlation between a graded loss of A20’s NF-κB inhibitory function 214



https://doi.org/10.1101/2020.02.23.961763


and an increasing hyper inflammatory in vivo phenotype with reduced phosphorylation at Ser381 215 13. The coding variant C243Y which showed the strongest in vivo phenotype and caused the 216

greatest reduction in Ser381 phosphorylation was originally identified in a four generation 217

Japanese family with hereditary autoinflammatory disease 7. In vitro biochemical studies show 218

that IκB kinase beta (IKKβ)-dependent phosphorylation at Ser381 is required for A20’s optimal 219

NF-κB and JNK suppressive function 13,20,43 (Figure 4B; Figure S5B and C). Both Ser381 and 220

Cys243 are conserved in zebrafish (Figure S2) allowing us to test the impact of alanine substitution 221

at Ser381, and the impact of the disease causing variant C243Y on the ability of human A20 to 222

rescue A20∆127/∆127 zebrafish from lethality. 223

224

As shown in Figure 4B, A20∆127/∆127 embryos that received either S381A or C243Y expressing 225

constructs were found in reduced numbers when compared to A20∆127/∆127 embryos that received 226

wild type hA20-expressing constructs. Specificity for the impact of S381A and C243Y on A20’s 227

in vivo function was shown by ectopic injection of another OTU domain mutant, namely the 228

C103A DUB mutant 13,21,22, which exhibited no negative impact on hA20’s ability to block NF-229

kB or JNK signalling in vitro (respectively Figure 4B; Figure S5B and C), nor A20’s ability to 230

rescue A20∆127/∆127 embryos from lethality (Figure 4A). The in vivo rescue properties of 231

individual A20 mutants did not relate to differences in protein stability, as shown by Western blot 232

analysis of A20 protein levels when expressed in cell lines (Figure 4B) 13, nor differences in the 233

levels of transgene expression in vivo as measured by the levels of GFP fluorescence which is co-234

expressed with A20 in the Ubb-hA20-P2A-EGFP construct (Figure S5A). 235



https://doi.org/10.1101/2020.02.23.961763


DISCUSSION 236

A20 inhibits NF-κB activation in mammals in response to TLR signals 1,9, and uncontrolled NF-237

κB activation results in excessive inflammation and lethality 44, yet there is little information 238

regarding the in vivo functional role of A20 beyond studies in mammalian models. An A20-like 239

protein sequence has been identified in the purple sea urchin and amphioxus 45 and in the lamprey 240

genomes 46 and here we identify a functional A20 homologue in zebrafish. In contrast, we did not 241

find a homologous A20 sequence in the genomes of lower invertebrates including Drosophila. 242

Similar to mammalian cells, and as shown here for zebrafish, A20 expression in amphioxus is also 243

regulated by LPS 19,45 and A20 modulates NF-κB signalling utilizing a ubiquitin editing function 244 13,21,45. However, the in vivo functional significance of A20 in amphioxus has not been tested 45. 245

Reminiscent of mouse data 1,9, we found that A20 deficiency in zebrafish was lethal and 246

remarkably re-introduction of human A20 could rescue a proportion of A20-deficient zebrafish 247

from lethality. We interpret these data to suggest that A20 arose during chordate evolution to 248

regulate inflammatory homeostasis and prevent lethality, most likely in response to environmental 249

microbial stimuli. As the major components of NF-κB signalling in mammals are conserved to 250

Drosophila 13,47 it is unclear as to why A20-dependent control of NF-κB signalling was required 251

for chordate evolution. Also, in mammalian cells A20 has been shown to regulate diverse 252

signalling cascades that control inflammation and cell death in both hematopoietic and non-253

hematopoietic cell lineages 10,48. Further investigation of A20’s function in non-mammalian 254

systems like zebrafish will reveal to what extent these critical mammalian functions extend to other 255

chordates and aid our understanding of the evolutionary significance of A20. 256

257

Gene targeting in mice to either introduce an alanine at Cys103 (C103A) that eliminates 258

biochemical OTU DUB activity 21, or a mutation that destroys the ZnF4 E3 ligase activity of A20 259 22, resulted in surprisingly little or no impact on lipopolysaccharide (LPS) responses and 260

inflammatory homeostasis in vivo 22,49, contrasting the lethal hyper-inflammatory phenotype of 261

A20-deficient mice 1,9. These findings indicate compensatory activity of the Cys103 and ZnF4 262

catalytic domains towards each other, and redundancy in the mechanism by which the OTU 263

domain regulates NF-κB activation which may include both catalytic and non-catalytic roles 12,14. 264

Previous biochemical studies have shown that A20’s NF-κB inhibitory activity requires 265

phosphorylation at Ser381 by IκB-kinase 20,21. We previously provided indirect evidence for the 266

importance of Ser381 in vivo 13. We have recently demonstrated how a series of A20 coding 267

variants that cause a decrease in phosphorylation at Ser381 also decrease A20’s NF-κB inhibitory 268

activity in a graded manner, with a corresponding loss in the ability of A20 to control inflammatory 269



https://doi.org/10.1101/2020.02.23.961763


homeostasis in vivo. Here, using a new zebrafish functional genomics approach, we show evidence 270

that Ser381 phosphorylation is needed for A20’s optimal anti-inflammatory function in vivo, as 271

targeted mutation of Ser381 impairs the ability of A20 to rescue A20∆127/∆127 zebrafish from 272

lethality. Prevention of Ser381 phosphorylation may disable both OTU DUB and ZnF4 ubiquitin 273

editing functions most likely explaining to the highly penetrant phenotype seen for both zebrafish 274

and mice 13 compared to the individual inactivation of catalytic domains 22,49. Further studies are 275

required to understand how phosphorylation at Ser381 impacts A20’s function but one possibility 276

is that phosphorylation causes changes to A20 structure facilitating interactions between A20’s 277

distinct ubiquitin binding and ubiquitin editing domains. 278

279

Mutations in the A20 ZnF7 non catalytic NF-κB inhibitory domain cause spontaneous 280

macrophage inflammasome activation and arthritis in mice 17. The cell type specific and different 281

phenotype of ZnF7 mutant mice to that reported for mice with reduced Ser381 phosphorylation 13 282

suggests that A20 regulatory functions can be distinguished from one another. Ser381 may regulate 283

both OTU and the ZnF4 dependent-ubiquitin editing functions, whilst ZnF7 regulates linear 284

ubiquitin chains outside the aforementioned mentioned domains. This highlights how specific A20 285

functional domains are important within particular physiological and/or cellular contexts. 286

Understanding how specific molecular changes in A20 function link to cellular phenotypes may 287

provide a pathway to understand the seemingly broad association of A20 to many different tissue 288

specific autoimmune diseases identified through GWAS 10 and the broad clinical phenotypes 289

emerging in patients identified with A20 haploinsufficiency 43,50,51. 290

291

A20 deletion results in severe autoinflammatory disease in humans 30,52-55. The severe 292

impact of mutating Ser381 on A20 function shown here, and the discovery of an ever increasing 293

number of new A20 coding variants through human sequencing studies 13, raise the question as to 294

the existence of further human coding variants which may present with highly penetrant 295

phenotypes and contribute to disease. The conservation of A20 in zebrafish allows us to construct 296

a functional genomics pipeline to investigate the impact of medically and scientifically important 297

A20 coding variants in vivo. The discovery 7 of the cysteine to tyrosine substitution at position 243 298

(C243Y) in a four generation Japanese family with a Behcet’s like disease, together with the high 299

level of sequence homology across the mammalian and zebrafish OTU domain which included 300

this residue, provided an opportunity to test the sensitivity of the zebrafish pipeline. In this 301

zebrafish model the C243Y mutation caused a reduction in A20’s ability to rescue homozygous 302



https://doi.org/10.1101/2020.02.23.961763


A20-null zebrafish from lethality. This in vivo proof of concept data in the zebrafish model 303

showing that C243Y is a reduction of function variant is supported by the human genetic data 7, 304

as well as biochemical and mouse data 13. A20 variants like C243Y that modify phosphorylation 305

at Ser381 therefore represent a novel class of variants with medically relevant disease-causing 306

potential. 307

308

We highlight the potential of the zebrafish system to advance our knowledge of medically 309

significant A20 gene variants by taking advantage of regions of high homology. In addition to 310

shedding light on the functional significance of A20 coding variants, the conservation of the 311

zebrafish A20 genomic locus may also aid the elucidation of non-coding variants. GWAS studies 312

associate A20 with multiple autoimmune conditions and many identified SNPs are non coding 10. 313

Analysis of conserved trans regulatory regions in the A20 locus marked by active methylation 314

marks revealed an overlap with two SNPs associated with Crohn’s disease (rs7753394; rs7773904) 315

(GRCh38) and one associated with Systemic lupus erythematosus (rs10499197) 56-58. As these 316

SNP’s lay in putative enhancer regions they may alter A20 expression levels thereby contributing 317

to disease. The development of a novel functional genomics pipeline that utilises the new A20-318

deficient zebrafish model provides a new tool to investigate the impact of TNFAIP3 genetic 319

variants in vivo. Understanding the underlying genetic code of A20 will have relevance for 320

understanding human disease mechanisms and the development of targeted drug therapies. This 321

same approach could be utilised to increase understanding of the impact of human genetic variation 322

for other highly conserved genes. 323

324



https://doi.org/10.1101/2020.02.23.961763


ACKNOWLEDGMENTS 325

We thank the Biological Testing Facility at the Garvan Institute of Medical Research for fish care, 326

and David Zahra (Genomic Engineering, Garvan Institute of Medical Research) for helpful advice 327

with design of TALEN constructs, cloning and site directed mutagenesis of A20 mutants. We 328

thank Dr John Rawls (Department of Cell and Molecular Physiology, University of North 329

Carolina, Chapel Hill) for the NF-κB:EGFP transgenic construct. D.C. and N.W.Z were each 330

supported by an Australian Postgraduate Award and N.W.Z. is an International Pancreas and Islet 331

Transplant Association (IPITA) Derek Gray Fellow. The research was supported by National 332

Health and Medical Research (NHMRC) grants to S.T.G. (1130222, 1189235) and K.K. 333

(1130247); as well as grants to S.T.G. from the NIH (DK076169) and the Australian Juvenile 334

Diabetes Research Foundation (3-SRA-2018-604-M-B). S.T.G. is NHMRC Senior Research 335

Fellow (1140691). 336

337

AUTHOR CONTRIBUTIONS 338

D.C., B.P., D.H. and S.T.G generated the Tg(NF-κB, c-FOS:EGFP)GI1 and D.C. and S.T.G. 339

generated the tnfaip3GI2 zebrafish lines. Genomic analysis of the zebrafish tnfaip3 locus was 340

conducted by O.B.. All in vivo zebrafish studies were conducted by D.C. Cloning of human A20 341

and A20 mutants, Western blotting experiments and NF-κB reporter assays were conducted by 342

D.C. and N.Z.W. JNK reporter assays were conducted by D.C., J.Z.M.H. and D.R.C.. J.B. and 343

T.C. carried out intravital imaging. Immunohistochemistry was conducted by J.W., D.C and E.S.. 344

Zebrafish expertise and essential experimental advice for use of zebrafish reporter lines was 345

provided by K.K. D.C., N.W.Z., B.P., J.Z.R.H., J.W., D.C., K.H., T.C., D.H. and S.T.G. critically 346

assessed data. D.C. and S.T.G. co-wrote the manuscript. S.T.G. devised and led the experimental 347

plan and is guarantor of the study. 348

349

DECLARATION OF INTERESTS 350

The authors declare no competing financial interests. 351



https://doi.org/10.1101/2020.02.23.961763


FIGURE LEGENDS. 352

Figure 1 A20-deficient zebrafish fail to thrive 353

A Zebrafish were stimulated with 75 µg/ml LPS and assayed for gene expression (mRNA) at 30 354

min intervals. Each data point represents an N=3 pools of 10 fish each per group from three 355

independent crosses. Data represented as mean ± SD. B Individual marker peak induction, and 356

relative p-values. *P < 0.05, **P < 0.01. C Gene targeting of A20 highlighting the 20 bp deletion 357

and predicted stop codon at amino acid 127, previously a Leucine. D Percentages of surviving 358

A20+/+, A20+/∆127 and A20∆127/∆127 zebrafish at 1, 2, 3 and 8wpf, generated from N=15 A20+/∆127 359

crosses. E Body area measured in (pixel/inch) per zebrafish genotype at 1 and 2 wpf. Each dot 360

represents an individual zebrafish. Data represented as mean ± SD, P-values were determined 361

using 1way ANOVA. *P < 0.05, ****P < 0.0001. Each group passed the normality test. F 362

Representative brightfield image of one A20+/+ (top) and A20∆127/∆127 (bottom) zebrafish at 2 wpf. 363

Scale bars represent 2 mm. G Body to liver ratios measured in (pixel/inch) per zebrafish genotype 364

at 1 and 2 wpf. Each dot represents an individual zebrafish. Data represented as mean ± SD, P-365

values were determined using 1way ANOVA. *P < 0.05, ****P < 0.0001. Each group passed the 366

normality test. For each genotype a representative A20+/+, A20+/∆127 and A20∆127/∆127 2 wpf 367

zebrafish is shown. The scale bars represent 2 mm. H Immuno-fluorescence sections of livers from 368

2CLIP reporter fish for the A20 genotypes. The scale bars represent 300µm. 369

370

Figure 2 A20 is necessary to maintain inflammatory homeostasis in zebrafish 371

A Representative fluorescent image of a NF-κB:eGFP zebrafish at 6 dpf showing spontaneous NF-372

κB activity (fluorescence) in the (a) mouth, (b) lateral neuronmasts, (c) pharyngeal tooth, and (d) 373

gills. Scale bar represents 2mm. B Representative fluorescent images of NF-κB:eGFP zebrafish of 374

different A20 genotypes at 3 and 6 dpf ± LPS. The scale bars represent 1mm. C-D Cumulative 375

data of relative fluorescence units as in (E) of NF-κB:eGFP zebrafish by A20 genotype at; (F) 3dpf 376

without LPS stimulation and, (G), 6 dpf ± LPS. Each dot represents an individual fish. Data are 377

representative of N=10 crosses. N per group and mean values reported below each graph. Data 378

represented as mean ± SD, P-values were determined using 1-way ANOVA. Each group passed 379

the normality test. **P < 0.01, ***P < 0.001, ****P < 0.0001. E Mpeg1 marker basal peaks at 380

1wpf, and relative p-values. *P < 0.05, N = 4 experiments of unstimulated zebrafish. 381

382



https://doi.org/10.1101/2020.02.23.961763


Figure 3 Impact of A20 deletion on macrophage activity in zebrafish. 383

A Representative fluorescent image of mpeg1.1:RFP;A20+/+ (top) and mpeg1.1:RFP;A20∆127/∆127 384

(bottom) zebrafish at 1wpf. B Representative fluorescent image of lck:RFP;A20+/+ (top) and 385

lck:RFP; A20∆127/∆127 (bottom) zebrafish at 1wpf. C Number of macrophages counted from 386

images of zebrafish by A20 genotype as in (A). Data represented as mean ± SD, P-values were 387

determined using Whelch’s t-test. **P < 0.01, ****P < 0.0001. D Number of T lymphocytes 388

counted from images of zebrafish by A20 genotype as in (B). Data represented as mean ± SD, P-389

values were determined using Whelch’s t-test. **P < 0.01, ****P < 0.0001. E Representative 390

transversal zebrafish head cryosections for a NF-κB:eGFP and mpeg1.1:RFP zebrafish for each 391

A20 genotype at 1 wpf. The scale bar represents 75 µm. F Average number of macrophages at the 392

site of injury measured every 2 hours and at 1 and 2 days post-injury. Error bands are SEM. Area 393

under the curve was significant for A20+/+ x A20+/∆127 (****P < 0.0001) and A20+/+ x A20∆127/∆127 394

(***P < 0.001). G Macrophages from mpeg1.1:RFP positive zebrafish were imaged using in vivo 395

two-photon microscopy and their track displacement, track duration and track length were 396

quantified. Data collected from 4 A20+/+ and 3 A20+/∆127 comprising of 68 and 98 macrophages 397

tracks respectively. Data represented as mean ± SD, P-values were determined using Mann-398

Whitney t-test. **P < 0.01, ****P < 0.0001. H Timed screenshots of a Two-Photon z-stack 15 399

minutes time-lapse of an mpeg1.1:RFP A20;A20+/+ and mpeg1.1:RFP A20;A20+/∆127 zebrafish. 400

The tracks show the path taken by a representative macrophage expressing RFP. The chosen 401

macrophages are the closest to the average tracking data. 402

403

Figure 4 The impact of human A20 mutants on A20∆127/∆127 zebrafish survival. 404

A Survival data for A20+/+, A20+/∆127 and A20∆127/∆127 zebrafish at 3 wpf with ectopic 405

administration of WT A20 (ref-hA20) or the C103A, S381A or C243Y mutants. Numbers indicate 406

number of genotyped survivors expressing the respective human gene variant. Each box indicates 407

the survivors from an independent experiment. Survival frequency data is represented in the pie 408

chart. N indicates the total number of zebrafish embryos injected for each A20 mutant. B NF-κB 409

luciferase assay for HEK293 cells transiently non-transfected (NTC), or co-transfected with WT 410

human-A20 or the C103A, S381A or C243Y mutant and treated with or without hTNFα. Data 411

represented as mean ± SD. Each dot represents an individual experiment and are representative of 412



https://doi.org/10.1101/2020.02.23.961763


N = 3-5 experiments per A20 mutant. Below is shown a representative Western blot for each A20 413

mutant. The displayed gel has been cropped to improve clarity and conciseness of the figure. 414

415



https://doi.org/10.1101/2020.02.23.961763


MATERIALS AND METHODS 416

Zebrafish 417

Fertilized eggs were collected from zebrafish (Danio rerio) group matings and incubated at 28.5 418

°C at a density <60 embryos in ~25 mL E3 egg water. Adult zebrafish were maintained under 419

standard conditions at 28.5 °C and pH 7-8 at a density of 5-10 fish L-1 under a photoperiodic cycle 420

of 14/10 light/dark hours. Genotyping was performed by fin clipping and HotSHOT DNA 421

extraction 59,60 followed by High Resolution Melting assay. Animal studies were approved by the 422

Garvan/St Vincent's Animal Ethics Committee. All procedures performed complied with the 423

Australian Code of Practice for Care and Use of Animals for Scientific Purposes. 424

425

Genomic structure of A20 426

HiC data (40Kb resolution) corresponding to H1 cells were visualized in the 3D genome browser 427 61 (hg19 genome assembly). Layered H3K27ac ChIP-seq data from seven cell lines (GM12878, 428

H1-hESC, HSMM, HUVEC, K562, NHEK, NHLF), available through the UCSC genome browser 429 62 were generated by the ENCODE consortium (2012). Zebrafish 24hpf embryo H3K27ac ChIP-430

seq reads were mapped by Bowtie 63 to unique genomic positions, followed by duplicate removal. 431

Mapped ChIP-seq reads were visualized in WIG format as previously described 64 utilising the 432

UCSC genome browser. 433

434

Tnfaip3 knockout line 435

A pair of TALENs targeting zebrafish tnfaip3 (A20) exon 2 were generated using “PLATINUM 436

Gate TALEN Kit” (Addgene, #1000000043). 0.2 ng of mRNA encoding the TALEN pair was 437

delivered into the cytoplasm of wild-type zebrafish embryos at the one-cell stage. Founders were 438

genotyped using the following primers: zfA20_Fw: AGTATCTGCTGGGGGTTCAGGA and 439

zfA20_Rev: CACCGCATGCAGAGCTT. Subsequently a founder line was identified harbouring 440

a 20 bp deletion in exon 2 with a predicted stop codon at amino acid 127. The tnfaip3dGI4 A20-441

deficient line was maintained in the heterozygous state (tnfaip3∆127/tnfaip3+) and crossed to the 442

zebrafish reporter lines described below. 443

444

Zebrafish reporter lines 445

The lck:RFP (TgBAC(lck:RFP)vcc4) and mpeg1.1:RFP (TgBAC(mpeg1.1:RFP)vcc7) reporter lines 446

have been described previously 36. The lymphocyte protein tyrosine kinase (lck) promoter targets 447

gene expression to T cells whereas macrophage expressed gene 1.1 (mpeg1.1) promoter targets 448

gene expression to macrophages. 449



https://doi.org/10.1101/2020.02.23.961763


The Tg(ins:dsRed)m1081;Tg(fabp10:dsRed;ela3l:GFP)gz12 line, known as 2-Color Liver Insulin 450

acinar Pancreas (2CLIP) fish line, has been described previously 65. This fish express dsRed 451

fluorescent protein both in the hepatocytes, driven by the fatty acid binding protein 10 (fabp10 452

gene), and in the islets of the endocrine pancreas, driven by the insulin promoter (insulin gene). 453

The exocrine pancreas instead expresses GFP driven by the pancreas specific transcriptor factor 454

1a (ptf1a gene). The Tg(NF-κB, c-FOS:EGFP)GI5 zebrafish line was generated using the Tol2 455

system to insert a DNA fragment containing six repeats of an NF-κB binding site in tandem with 456

a minimal c-Fos promoter that drive the expression of eGFP, followed by an SV40 polyadenylation 457

signal sequence (SV40pA) 24. 458

459

Fluorescence quantification through ImageJ 460

Live zebrafish were placed in a gel solution (1x E3 solution, 3% methyl cellulose, 5% Tricaine), 461

in a 150 mm petri dish lid. Fish were placed on the left side up. Images were processed using 462

ImageJ (v.1.50b). In the case of the NF-κB:eGFP line fluorescence measurements, an outline of 463

the fish was drawn by hand, excluding the yolk sack due to its auto fluorescence. Measurement 464

values of area (pixel/inch) and integrated density (area × mean grey value) were recorded for each 465

of the measured fish. Background value was subtracted to the fish intensity value. 466

467

NF-κB and JNK reporter assays 468

The NF-κB reporter assay was conducted exactly as described previously 19. In brief, human 469

Embryonic Kidney 293T (HEK293T) cells 66 were cultured in Dulbecco's Modified 470

Eagle's medium (DMEM) and incubated at 37°C in 5% CO2 and kept at a passage number between 471

21-40 for experiments. Cells were transfected using Lipofectamine 2000 (ThermoScientific 472

#11668019) with β-galactosidase (β-gal) and NF-κB.Luc with either ‘empty’ pcDNA3.1 vector 473

(NTC; non template control) or pcDNA3.1 encoding wild-type human-A20 or either the A20 474

mutant Cys103Ala or Ser38Ala. HEK293T cells were stimulated with 300 U/ml of recombinant 475

human TNFα (RandD Systems, Minneapolis, MN; #210-TA) for 8 hours. To assess the effect of 476

A20 mutants on JNK signalling, HEK293T cells which constitutively expresses mCherry Kinase 477

Translocation Reporter (mCherryKTR) which translocates from the nucleus to the cytoplasm when 478

phosphorylated by JNK were used (de la Cova, Townley et al. 2017). JNK activity was determined 479

by measuring the ratio between nucleus and cytoplasmic mCherry in cells transiently transfected 480

with human-A20 mutants (using JetPRIME reagent) following human TNFα stimulation. Each 481

experiment included automated cell imaging (Thermo Scientific Cellomics Arrayscan machine) of 482



https://doi.org/10.1101/2020.02.23.961763


~150, 000 to ~350, 000 individual cells per well in duplicate wells over a period of 60 minutes (5 483

minutes intervals). 484

485

Generation of A20 mutants for ectopic expression in zebrafish 486

Two A20 mutants were generated for this study. Mutations were introduced into cDNA encoding 487

full length human A20 by site directed mutagenesis to cause an alanine substitution at position 488

Cys103 (C103A) and at Ser381 (S381A). Wild type A20 or either mutant C103A or S381A A20-489

cDNA were cloned into the expression cassette Ubb-[A20 insert]-P2A-EGFP and ~100 ng/µl of 490

DNA was co-injected with I-SceI meganuclease to promote random integration in the zebrafish 491

genome. Glass needles for zebrafish eggs injections were prepared from glass capillary tubes with 492

filament (ø1 mm × 90 mm) (WPI #1B100F-3) using a NARISHIGE dual-stage micropipette puller 493

PC-10 machine. The injections were all performed between the one and two-cell stage of 494

development. 495

496

Two-photon intravital microscopy 497

Two-photon imaging was performed using an upright Zeiss 7MP two-photon microscope (Carl 498

Zeiss) with a W Plan-Apochromat 20×/1.0 DIC (UV) Vis-IR water immersion objective. Four 499

external NDDs were used to detect blue (SP 485), green (BP 500-550), red (BP 565-610) and far 500

red (BP 640-710). High repetition rate femtosecond pulsed excitation was provided by a 501

Chameleon Vision II Ti:Sa laser (Coherent Scientific) with 690-1064nm tuning range. We 502

acquired 3µm z-steps at 512×512 pixels and resolution 0.83µm/pixel at a frame rate of 10 fps and 503

dwell time of 1.27 µs/pixel using bidirectional scanning. Intravital two-photon microscopy was 504

performed by embedding 1wpf zebrafish in 1% low melting point agarose in E3 water and 0.4% 505

Tricaine to limit the overall tissue drifting and motility of the fish. 506

507

Image processing and data analysis 508

Raw image files were processed using Imaris (Bitplane) software. A Gaussian filter was applied 509

to reduce background noise. Tracking was performed using Imaris spot detection function to locate 510

the centroid of cells. Motility parameters such as cell displacement (or track length calculated as 511

the total length of displacements within the track) were obtained using Imaris Statistics function. 512

513

Statistical analysis 514



https://doi.org/10.1101/2020.02.23.961763


Data were analysed as appropriate using a nonparametric Mann-Whitney test, a Whelch t-test, 515

Area under the curve (AUC) or a one-way ANOVA test. Statistical analyses were carried out using 516

GraphPad Prism v7.0. In all cases, the significance threshold was set at p £ 0.05. 517

518



https://doi.org/10.1101/2020.02.23.961763


ONLINE SUPPLEMENTAL MATERIAL 519

520

Figure S1. Genomic structure of the TNFAIP3 (A20) locus. 521

A HiC data (H1 cells) visualized over the A20 locus, overlapping A20 SNPs and H3K27ac ChIP-522

seq signal. Insets correspond to three evolutionarily-conserved (PhyloP, vertebrate conservation) 523

A20 SNPs, which coincide with the regulatory H3K27ac mark. The alternating yellow and pink 524

shaded areas are predicted Topologically associated domains (TADs), whereas the heatmap 525

(white-red) represents normalised interaction frequency, as described in (Wang, Song et al. 2018). 526

The SNP IDs for SNPs with insets in the figure are (from left to right on the figure): a) 527

chr6:138085249-138085249 (rs7753394), b) chr6:138085366-138085366 (rs7773904), c) 528

chr6:138132517-138132517 (rs10499197). B H3K27ac ChIP-seq (24 hpf embryos) visualized 529

over the A20 locus. Dark boxes represent human-fish DNA sequence conservation. 530

531

Figure S2. Protein homology of TNFAIP3 across multiple vertebrate lineages 532

The consensus sequence (cons) below each amino acid reports an “*” (asterisk) when all the amino 533

acids are the same, a “:” (colon) indicates conservation between groups of strongly similar 534

properties and a “.” (period) indicates conservation between groups of weakly similar properties. 535

The amino acids comprising the OTU domain catalytic triad (Asp70, Cys103, His256) and position 536

Ser381 are also shown. Note these two protein motifs are explicitly conserved across species. 537

538

Figure S3. Validation of A20 deletion in TALEN targeted zebrafish 539

A HRMA curves (derivative on the left and raw-melt on the right) showing the three possible 540

genotypes arising from a cross between two A20+/∆127. The trace for a zebrafish carrying either 541

A20 wild type alleles (Blue), or heterozygous (Green) or homozygous (Red) for the mutant A20 542

allele are shown. B Shows sequence trace and alignment for a zebrafish carrying either the A20 543

wild type A20 allele or the deletion. C Shows PCR amplification of the TALEN targeted region 544

of the A20 gene for A20+/+, A20+/∆127 and A20∆127/∆127 zebrafish. The displayed gel has been 545

cropped to improve clarity and conciseness of the figure. 546

547

Figure S4. Co-localisation of NF-κB and MPEG1 reporter signals in zebrafish 548

Two-photon microscopy 3D-volume of a double reporter (NF-κB:eGFP;mpeg1.1:RFP) A20+/+. 549

Below are shown each individual channel and the merge to demonstrate colocalization of the two 550



https://doi.org/10.1101/2020.02.23.961763


markers and highlighted by dotted white lines circling the cells where colocalization occurrs. 551

Image supplemented by video file (Video S3). 552

553

Figure S5. Control Data for A20-dependent JNK inhibition and in vivo gene expression 554

A Relative fluorescence units, corrected for fish size. No statistical significance was observed 555

between groups of fish injected with different human A20-EGFP constructs. B Impact of A20 556

variants in mammalian cells. JNK activation kinetics for HEK293 cells transiently non-transfected 557

(NTC), or co-transfected with WT human-A20 or the C103A, S381A or C243Y mutant and treated 558

with or without hTNFα for 1 hour and quantified at 5 minutes intervals. Data shows the ratio of 559

Cytoplasm over Nuclear (C/N) JNK activation, represented as mean ± standard error. C Data from 560

the 15 min time point for all A20 variants is represented in heat map format. Statistical analysis 561

carried as Area Under the Curve (AUC) compared to WT hA20. WT and NTC = ****P < 0.0001, 562

WT and S381A or C243Y = *P < 0.05. The data on this graph have been generated by three 563

independent experiments. 564

565



https://doi.org/10.1101/2020.02.23.961763


REFERENCES 566

567

568

569

570

1 Lee, E. G. Failure to Regulate TNF-Induced NF-kappa B and Cell Death 571

Responses in A20-Deficient Mice. Science 289, 2350-2354, 572

doi:10.1126/science.289.5488.2350 (2000). 573

2 Cooper, J. T. et al. A20 blocks endothelial cell activation through a NF-kappaB-574

dependent mechanism. J. Biol. Chem. 271, 18068-18073 (1996). 575

3 Grey, S. T., Arvelo, M. B., Hasenkamp, W., Bach, F. H. & Ferran, C. A20 inhibits 576

cytokine-induced apoptosis and nuclear factor kappaB-dependent gene activation 577

in islets. J. Exp. Med. 190, 1135-1146 (1999). 578

4 Song, H. Y., Rothe, M. & Goeddel, D. V. The tumor necrosis factor-inducible 579

zinc finger protein A20 interacts with TRAF1/TRAF2 and inhibits NF-kappaB 580

activation. Proc. Natl. Acad. Sci. U. S. A. 93, 6721-6725 (1996). 581

5 Jaattela, M., Mouritzen, H., Elling, F. & Bastholm, L. A20 zinc finger protein 582

inhibits TNF and IL-1 signaling. J. Immunol. 156, 1166-1173 (1996). 583

6 Dixit, V. M. et al. Tumor necrosis factor-alpha induction of novel gene products 584

in human endothelial cells including a macrophage-specific chemotaxin. J. Biol. 585

Chem. 265, 2973-2978 (1990). 586

7 Shigemura, T. et al. Novel heterozygous C243Y A20/TNFAIP3 gene mutation is 587

responsible for chronic inflammation in autosomal-dominant Behcet's disease. 588

RMD Open 2, e000223, doi:10.1136/rmdopen-2015-000223 (2016). 589

8 Zhu, L. et al. Characteristics of A20 gene polymorphisms and clinical significance 590

in patients with rheumatoid arthritis. J. Transl. Med. 13, 215, doi:10.1186/s12967-591

015-0566-1 (2015). 592



https://doi.org/10.1101/2020.02.23.961763


9 Boone, D. L. et al. The ubiquitin-modifying enzyme A20 is required for 593

termination of Toll-like receptor responses. Nat. Immunol. 5, 1052-1060, 594

doi:10.1038/ni1110 (2004). 595

10 Ma, A. & Malynn, B. A. A20: linking a complex regulator of ubiquitylation to 596

immunity and human disease. Nat. Rev. Immunol. 12, 774-785, 597

doi:10.1038/nri3313 (2012). 598

11 Vereecke, L., Beyaert, R. & van Loo, G. Genetic relationships between 599

A20/TNFAIP3, chronic inflammation and autoimmune disease. Biochem. Soc. 600

Trans. 39, 1086-1091, doi:10.1042/bst0391086 (2011). 601

12 Wertz, I. E. et al. De-ubiquitination and ubiquitin ligase domains of A20 602

downregulate NF-kappaB signalling. Nature 430, 694-699, 603

doi:10.1038/nature02794 (2004). 604

13 Zammit, N. W. et al. Denisovan, modern human and mouse TNFAIP3 alleles tune 605

A20 phosphorylation and immunity. Nat. Immunol. 20, 1299-1310, 606

doi:10.1038/s41590-019-0492-0 (2019). 607

14 Skaug, B. et al. Direct, noncatalytic mechanism of IKK inhibition by A20. Mol. 608

Cell 44, 559-571, doi:10.1016/j.molcel.2011.09.015 (2011). 609

15 Evans, P. C. et al. Zinc-finger protein A20, a regulator of inflammation and cell 610

survival, has de-ubiquitinating activity. Biochem. J. 378, 727-734, 611

doi:10.1042/bj20031377 (2004). 612

16 Shembade, N., Ma, A. & Harhaj, E. W. Inhibition of NF-kappaB signaling by A20 613

through disruption of ubiquitin enzyme complexes. Science 327, 1135-1139, 614

doi:10.1126/science.1182364 (2010). 615

17 Polykratis, A. et al. A20 prevents inflammasome-dependent arthritis by inhibiting 616

macrophage necroptosis through its ZnF7 ubiquitin-binding domain. Nat. Cell 617

Biol. 21, 731-742, doi:10.1038/s41556-019-0324-3 (2019). 618



https://doi.org/10.1101/2020.02.23.961763


18 Krikos, A., Laherty, C. D. & Dixit, V. M. Transcriptional activation of the tumor 619

necrosis factor alpha-inducible zinc finger protein, A20, is mediated by kappa B 620

elements. J. Biol. Chem. 267, 17971-17976 (1992). 621

19 Liuwantara, D. et al. Nuclear factor-kappaB regulates beta-cell death: a critical 622

role for A20 in beta-cell protection. Diabetes 55, 2491-2501, doi:10.2337/db06-623

0142 (2006). 624

20 Hutti, J. E. et al. IkappaB kinase beta phosphorylates the K63 deubiquitinase A20 625

to cause feedback inhibition of the NF-kappaB pathway. Mol. Cell. Biol. 27, 7451-626

7461, doi:10.1128/mcb.01101-07 (2007). 627

21 Wertz, I. E. et al. Phosphorylation and linear ubiquitin direct A20 inhibition of 628

inflammation. Nature 528, 370-375, doi:10.1038/nature16165 (2015). 629

22 Lu, T. T. et al. Dimerization and ubiquitin mediated recruitment of A20, a 630

complex deubiquitinating enzyme. Immunity 38, 896-905, 631

doi:10.1016/j.immuni.2013.03.008 (2013). 632

23 Correa, R. G. et al. Characterization of NF-kappa B/I kappa B proteins in zebra 633

fish and their involvement in notochord development. Mol. Cell. Biol. 24, 5257-634

5268, doi:10.1128/mcb.24.12.5257-5268.2004 (2004). 635

24 Kanther, M. et al. Microbial Colonization Induces Dynamic Temporal and Spatial 636

Patterns of NF-κB Activation in the Zebrafish Digestive Tract. Gastroenterology 637

141, 197-207, doi:http://dx.doi.org/10.1053/j.gastro.2011.03.042 (2011). 638

25 Novoa, B., Bowman, T. V., Zon, L. & Figueras, A. LPS response and tolerance 639

in the zebrafish (Danio rerio). Fish Shellfish Immunol 26, 326-331, 640

doi:10.1016/j.fsi.2008.12.004 (2009). 641

26 Creyghton, M. P. et al. Histone H3K27ac separates active from poised enhancers 642

and predicts developmental state. Proc. Natl. Acad. Sci. U. S. A. 107, 21931-643

21936, doi:10.1073/pnas.1016071107 (2010). 644



https://doi.org/10.1101/2020.02.23.961763


27 Rada-Iglesias, A. et al. A unique chromatin signature uncovers early 645

developmental enhancers in humans. Nature 470, 279-283, 646

doi:10.1038/nature09692 (2011). 647

28 Komander, D. & Barford, D. Structure of the A20 OTU domain and mechanistic 648

insights into deubiquitination. Biochem. J. 409, 77-85, doi:10.1042/BJ20071399 649

(2008). 650

29 Phelan, P. E., Mellon, M. T. & Kim, C. H. Functional characterization of full-651

length TLR3, IRAK-4, and TRAF6 in zebrafish (Danio rerio). Mol. Immunol. 42, 652

1057-1071, doi:10.1016/j.molimm.2004.11.005 (2005). 653

30 Matmati, M. et al. A20 (TNFAIP3) deficiency in myeloid cells triggers erosive 654

polyarthritis resembling rheumatoid arthritis. Nat. Genet. 43, 908-912, 655

doi:10.1038/ng.874 (2011). 656

31 Her, G. M., Chiang, C. C., Chen, W. Y. & Wu, J. L. In vivo studies of liver-type 657

fatty acid binding protein (L-FABP) gene expression in liver of transgenic 658

zebrafish (Danio rerio). FEBS Lett. 538, 125-133 (2003). 659

32 Ellett, F., Pase, L., Hayman, J. W., Andrianopoulos, A. & Lieschke, G. J. mpeg1 660

promoter transgenes direct macrophage-lineage expression in zebrafish. Blood 661

117, e49-56, doi:10.1182/blood-2010-10-314120 (2011). 662

33 Wittamer, V., Bertrand, J. Y., Gutschow, P. W. & Traver, D. Characterization of 663

the mononuclear phagocyte system in zebrafish. Blood 117, 7126-7135, 664

doi:10.1182/blood-2010-11-321448 (2011). 665

34 Benard, E. L. et al. Macrophage-expressed perforins mpeg1 and mpeg1.2 have an 666

anti-bacterial function in zebrafish. J. Innate Immun. 7, 136-152, 667

doi:10.1159/000366103 (2015). 668

35 Herbomel, P., Thisse, B. & Thisse, C. Ontogeny and behaviour of early 669

macrophages in the zebrafish embryo. Development 126, 3735-3745 (1999). 670



https://doi.org/10.1101/2020.02.23.961763


36 Sugimoto, K., Hui, S. P., Sheng, D. Z., Nakayama, M. & Kikuchi, K. Zebrafish 671

FOXP3 is required for the maintenance of immune tolerance. Dev. Comp. 672

Immunol. 73, 156-162, doi:10.1016/j.dci.2017.03.023 (2017). 673

37 Kigerl, K. A. et al. Identification of two distinct macrophage subsets with 674

divergent effects causing either neurotoxicity or regeneration in the injured mouse 675

spinal cord. J. Neurosci. 29, 13435-13444, doi:10.1523/jneurosci.3257-09.2009 676

(2009). 677

38 Jablonski, K. A. et al. Novel Markers to Delineate Murine M1 and M2 678

Macrophages. PLoS ONE 10, e0145342, doi:10.1371/journal.pone.0145342 679

(2015). 680

39 Renshaw, S. A. et al. A transgenic zebrafish model of neutrophilic inflammation. 681

Blood 108, 3976-3978, doi:10.1182/blood-2006-05-024075 (2006). 682

40 Grabher, C. & Wittbrodt, J. Meganuclease and transposon mediated transgenesis 683

in medaka. Genome Biol. 8, 1-7, doi:10.1186/gb-2007-8-s1-s10 (2007). 684

41 Soroldoni, D., Hogan, B. M. & Oates, A. C. in Zebrafish: Methods and Protocols 685

(eds J. Graham Lieschke, C. Andrew Oates, & Koichi Kawakami) 117-130 686

(Humana Press, 2009). 687

42 Hoshijima, K., Jurynec, Michael J. & Grunwald, David J. Precise Editing of the 688

Zebrafish Genome Made Simple and Efficient. Dev. Cell 36, 654-667, 689

doi:10.1016/j.devcel.2016.02.015 (2016). 690

43 Zhou, Q. et al. Loss-of-function mutations in TNFAIP3 leading to A20 691

haploinsufficiency cause an early-onset autoinflammatory disease. Nat. Genet. 692

48, 67-73, doi:10.1038/ng.3459 (2016). 693

44 Beg, A. A., Sha, W. C., Bronson, R. T., Ghosh, S. & Baltimore, D. Embryonic 694

lethality and liver degeneration in mice lacking the RelA component of NF-kappa 695

B. Nature 376, 167-170, doi:10.1038/376167a0 (1995). 696



https://doi.org/10.1101/2020.02.23.961763


45 Yuan, S. et al. Emergence of the A20/ABIN-mediated inhibition of NF-kappaB 697

signaling via modifying the ubiquitinated proteins in a basal chordate. Proc. Natl. 698

Acad. Sci. U. S. A. 111, 6720-6725, doi:10.1073/pnas.1321187111 (2014). 699

46 Smith, J. J. et al. The sea lamprey germline genome provides insights into 700

programmed genome rearrangement and vertebrate evolution. Nat. Genet. 50, 701

270-277, doi:10.1038/s41588-017-0036-1 (2018). 702

47 Dushay, M. S., Asling, B. & Hultmark, D. Origins of immunity: Relish, a 703

compound Rel-like gene in the antibacterial defense of Drosophila. Proc. Natl. 704

Acad. Sci. U. S. A. 93, 10343-10347, doi:10.1073/pnas.93.19.10343 (1996). 705

48 Catrysse, L., Vereecke, L., Beyaert, R. & van Loo, G. A20 in inflammation and 706

autoimmunity. Trends Immunol. 35, 22-31, doi:10.1016/j.it.2013.10.005 (2014). 707

49 De, A., Dainichi, T., Rathinam, C. V. & Ghosh, S. The deubiquitinase activity of 708

A20 is dispensable for NF-kappaB signaling. EMBO Rep. 15, 775-783, 709

doi:10.15252/embr.201338305 (2014). 710

50 Duncan, C. J. A. et al. Early-onset autoimmune disease due to a heterozygous 711

loss-of-function mutation in TNFAIP3 (A20). Ann. Rheum. Dis., 712

doi:10.1136/annrheumdis-2016-210944 (2017). 713

51 Li, G. M. et al. Expanding the spectrum of A20 haploinsufficiency in two Chinese 714

families: cases report. BMC Med Genet 20, 124, doi:10.1186/s12881-019-0856-1 715

(2019). 716

52 Chu, Y. et al. B cells lacking the tumor suppressor TNFAIP3/A20 display 717

impaired differentiation and hyperactivation and cause inflammation and 718

autoimmunity in aged mice. Blood 117, 2227-2236, doi:10.1182/blood-2010-09-719

306019 (2011). 720

53 Heger, K. et al. A20-deficient mast cells exacerbate inflammatory responses in 721

vivo. PLoS Biol. 12, e1001762, doi:10.1371/journal.pbio.1001762 (2014). 722



https://doi.org/10.1101/2020.02.23.961763


54 Nagamachi, A. et al. Acquired deficiency of A20 results in rapid apoptosis, 723

systemic inflammation, and abnormal hematopoietic stem cell function. PLoS 724

ONE 9, e87425, doi:10.1371/journal.pone.0087425 (2014). 725

55 Nakagawa, M. M., Thummar, K., Mandelbaum, J., Pasqualucci, L. & Rathinam, 726

C. V. Lack of the ubiquitin-editing enzyme A20 results in loss of hematopoietic 727

stem cell quiescence. J. Exp. Med., doi:10.1084/jem.20132544 (2015). 728

56 Eleftherohorinou, H. et al. Pathway analysis of GWAS provides new insights into 729

genetic susceptibility to 3 inflammatory diseases. PLoS ONE 4, e8068, 730

doi:10.1371/journal.pone.0008068 (2009). 731

57 Vereecke, L. et al. A20 controls intestinal homeostasis through cell-specific 732

activities. Nat. Commun. 5, 5103, doi:10.1038/ncomms6103 (2014). 733

58 Graham, R. R. et al. Genetic variants near TNFAIP3 on 6q23 are associated with 734

systemic lupus erythematosus. Nat. Genet. 40, 1059-1061, doi:10.1038/ng.200 735

(2008). 736

59 Meeker, N. D., Hutchinson, S. A., Ho, L. & Trede, N. S. Method for isolation of 737

PCR-ready genomic DNA from zebrafish tissues. Biotechniques 43, 610, 612, 614 738

(2007). 739

60 Thomsen, N., Ali, R. G., Ahmed, J. N. & Arkell, R. M. High resolution melt 740

analysis (HRMA); a viable alternative to agarose gel electrophoresis for mouse 741

genotyping. PLoS ONE 7, e45252, doi:10.1371/journal.pone.0045252 (2012). 742

61 Wang, Y. et al. The 3D Genome Browser: a web-based browser for visualizing 743

3D genome organization and long-range chromatin interactions. Genome Biol. 19, 744

151, doi:10.1186/s13059-018-1519-9 (2018). 745

62 Kent, W. J. et al. The human genome browser at UCSC. Genome Res. 12, 996-746

1006, doi:10.1101/gr.229102 (2002). 747

63 Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-748

efficient alignment of short DNA sequences to the human genome. Genome Biol. 749

10, R25, doi:10.1186/gb-2009-10-3-r25 (2009). 750



https://doi.org/10.1101/2020.02.23.961763


64 Bogdanovic, O. et al. Dynamics of enhancer chromatin signatures mark the 751

transition from pluripotency to cell specification during embryogenesis. Genome 752

Res. 22, 2043-2053, doi:10.1101/gr.134833.111 (2012). 753

65 Farooq, M. et al. Histone deacetylase 3 (hdac3) is specifically required for liver 754

development in zebrafish. Dev. Biol. 317, 336-353, 755

doi:10.1016/j.ydbio.2008.02.034 (2008). 756

66 Graham, F. L., Smiley, J., Russell, W. C. & Nairn, R. Characteristics of a human 757

cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 36, 758

59-74, doi:10.1099/0022-1317-36-1-59 (1977). 759

760



https://doi.org/10.1101/2020.02.23.961763


A B

C

D E

F G

0

25

50

75

100

125

150

175

200

Rel

ativ

e flu

ores

cenc

e un

its (R

FU) ***

**

N = 46 66 34

Mean RFU = 74.31 91.11 100.7

0

100

200

300

400

500

600

700

800

Rel

ativ

e flu

ores

cenc

e un

its (R

FU)

LPS N = 11 32 33 32 19 18

***

********

****

A20∆127/∆127;NF-κB:EGFP

A20+/∆127;NF-κB:EGFP

A20+/+;NF-κB:EGFP

Mean RFU = 227.6 268.8 226.3 297.8 296.4 401.3

- + - + - +- - - LPS

a

a

b

b

c

c

d

d

A20+/+;NF-κB:EGFP

3dpf

6dpf

NO

LP

S

LP

S

[75 μ

g/m

l]

A20∆127/∆127;NF-κB:EGFPA20+/∆127;NF-κB:EGFPA20+/+;NF-κB:EGFP

A20+/+

A20∆127/∆127

TALEN target site

OTU 1 2 3 4 5

GGGAATTGCCTGCTTCACGCAGCGTCTCAGTATCTGCTGGGGGTTCAG- - - - - - - - - - - - - - - - - - - - - - - GAAAGCTCTGCATGCGGTGCTGAAGGAAACGGACACGACTGATTTCCT*

GGGAATTGCCTGCTTCACGCAGCGTCTCAGTATCTGCTGGGGGTTCAGGACACAGATCTGGTGCTGAGGAAAGCTCTGCATGCGGTGCTGAAGGAAACGGACACGACTGATTTCCT

0h 30min0.0

0.5

1.0

1.5

2.0

2.5

Rela

tive

expr

essio

n to

β-a

ctin

*

Rela

tive

expr

essio

n to

β-a

ctin

0h 30min0

1

2

3

4

5 **

0h 1h0

1

2

3

4

5 *

0h 1h0

2

4

6

8

10 **

zf A20 zf TNFα zf IL1β zf IL6

Rela

tive

expr

essio

n to

β-a

ctin

Rela

tive

expr

essio

n to

β-a

ctin

Rela

tive

expr

essio

n to

β-a

ctin

Rela

tive

expr

essio

n to

β-a

ctin

0 0.5 1 1.5 2 40123456789

10

zf A20

zf IL1βzf IL-6

zf TNFα

Time post LPS stimulation (hours)



https://doi.org/10.1101/2020.02.23.961763


B C

D E

A

10

20

30

40

50

60

70

(Body a

rea/L

iver

are

a)

***

1 2Weeks post-fertilization

***

*

3

4

5

6

7

8

9

10

Body a

rea (

pix

el/i

nch)

****

1 2Weeks post-fertilization

****

DA

PI

dsR

ED

Ca

sp

ase

-3

A20+/+;2CLIP A20+/∆127;2CLIP A20∆127/∆127;2CLIP

A20+/+;2CLIP

A20+/∆127;2CLIP

A20∆127/∆127;2CLIP

1 2 3 80

20

40

60

80

100

Age (weeks)

Observ

ed g

enoty

pe % 24.5

52.5

23

12.9

55.8

31.3

56.5

43.5

59.1

40.9

N = 322 - 294 - 62 - 22

A20∆127/∆127

A20+/∆127

A20+/+

A20∆127/∆127

A20+/∆127

A20+/+

A20∆127/∆127

A20+/+

A20∆127/∆127

A20+/∆127

A20+/+



https://doi.org/10.1101/2020.02.23.961763


lck:R

FP

mpeg1.1

:RF

PN

F-ĸ

B:E

GF

P

A20∆127/∆127A20+/∆127A20+/+

DA

PI

0

10

20

30

1 week post-fertilization

T-c

ells

/ b

ody s

ize r

atio

*******

N = 20 34 23

A20+/+;lck:RFP

A20+/∆127;lck:RFP

A20∆127/∆127;lck:RFP

0

10

20

30

1 week post-fertilization

Macro

phages / b

ody s

ize r

atio A20+/+;mpeg1.1:RFP

A20+/∆127;mpeg1.1:RFP

A20∆127/∆127;mpeg1.1:RFP

********

N = 21 38 19

A20+/+;mpeg1.1:RFP

A20+/∆127;mpeg1.1:RFP

µm

/s

N = 68 98

A20+/+;mpeg1.1:RFP

A20∆127/∆127;mpeg1.1:RFP

A20+/+;lck:RFP

A20∆127/∆127;lck:RFP

A B

E F

G H

C D

mpeg1.1:RFP

NF-ĸB:EGFP

Collagen:SHG

COLOCALIZATION

0.00

0.05

0.10

0.15

0.20

0.25****

mpeg1.1:RFPNF-ĸB:EGFPCollagen:SHG A20+/+

mpeg1.1:RFPNF-ĸB:EGFP Colocalization



https://doi.org/10.1101/2020.02.23.961763


A B

C D

E

A20 (90 kDa)

β-actin (42 kDa)

0

100

200

300

400

500

600

NF

-κB

relat

ive flu

ores

cenc

e un

its

NTC hA20WT C103A S381A

hTNFα

****

- + - + - + - +

0 5 10 15 20 25 30 35 400.4

0.6

0.8

1.0

1.2

Time post-stimulation (min)

C/N

ratio

NTC

hA20WT

S381A

C103A

15 m

in

NTC

S381A

C103A

hA20WT

Low

High

JNK activatio

n

hA20WT C103A S381A-

A20+/+ = 27

A20+/∆127 = 35

A20∆127/∆127 = 0

A20+/+ = 9

A20+/∆127 = 13

A20∆127/∆127 = 0

A20+/+ = 7

A20+/∆127 = 9

A20∆127/∆127 = 3

A20+/+ = 13

A20+/∆127 = 24

A20∆127/∆127 = 4

A20+/+ = 16

A20+/∆127 = 21

A20∆127/∆127 = 3

A20+/+ = 24

A20+/∆127 = 47

A20∆127/∆127 = 5

A20+/+ = 5

A20+/∆127 = 4

A20∆127/∆127 = 1

A20+/+ = 7

A20+/∆127 = 7

A20∆127/∆127 = 2

A20+/+ = 18

A20+/∆127 = 25

A20∆127/∆127 = 2

A20+/+ = 7

A20+/∆127 = 23

A20∆127/∆127 = 8

A20+/+ = 5

A20+/∆127 = 24

A20∆127/∆127 = 0

A20+/+ = 8

A20+/∆127 = 28

A20∆127/∆127 = 4

A20+/+ = 18

A20+/∆127 = 26

A20∆127/∆127 = 1

A20+/+ = 16

A20+/∆127 = 26

A20∆127/∆127 = 1

3.64% A20+/+

A20+/∆127

A20∆127/∆127

N = 162 176 109 157

A20+/+ = 20

A20+/∆127 = 34

A20∆127/∆127 = 0

A20+/+ = 8

A20+/∆127 = 16

A20∆127/∆127 = 0

37.96%

62.04% 55.06%

35.03%

9.91%

49.96%

38.04%

12%

67.75%

28.61%

40

100

100

130

100

130

hA

20

WT

A20 (Ms)

P-S381-A20 (Rb)

IκKβ

β-actin

S381A

PCDNA

- hTNFα + hTNFα

DA

PI

P6

5

hA20WT

- hTNFα + hTNFα

C103A

- hTNFα + hTNFα

S381A

- hTNFα + hTNFα



https://doi.org/10.1101/2020.02.23.961763




https://doi.org/10.1101/2020.02.23.961763


0

10

20

30

40

m

TRACK DISPLACEMENT

**** A20+/+;mpeg1.1:RFP

A20+/∆127;mpeg1.1:RFP

0

20

40

60

80

100

m

TRACK LENGTH

**** A20+/+;mpeg1.1:RFP

0

200

400

600

800

1000

seconds

TRACK DURATION

** A20+/+;mpeg1.1:RFP

A20+/∆127;mpeg1.1:RFP

A

B

C

A20+/∆127;mpeg1.1:RFP



https://doi.org/10.1101/2020.02.23.961763


a zebrafish functional genomics model to …...2020/02/23 · 88 many of the mammalian components...

Documents