HTP and UHTP methods for protein production and interactions - P4EU [PDF]

HTP and UHTP methods for protein production and interactions R. Vincentelli [email protected]

Manager and method development, HTP protein production facility AFMB UMR 7257, Marseille, France

“Automated” Structural Biology core facility opened to customers since 2004 (set up started in 2001 on SG project) Expression screen Cloning 6XHis

Soluble

Large scale expression

Insoluble Refolding screening

Target

E. Coli Prot-DNA Prot-prot

Design/validation of new functional assays - In vitro prot-prot interaction: HTP Pull down - In vitro Prot-DNA interaction: HTP SELEX

Mammalian HEK293, CHO Insect cells Baculovirus and Schneider cells

6.-*7('8*9

Nanodrop crystallization

Purification

:;9'7%0/ 600/week *Custom/private Entry PCR clone collections (pENTR) Entry vectors (pENTR) *Synthetic genes

Sequencing

Destination vectors (pDONR)

(no cloning since 2012)

collaborators Multiple hosts/systems

E. coli Multiple Tag: His, Strep, MBP, TRX, GST, GFP… (20) 6XHis

Tag attB1 protein

Tev

Target

attB2

Protein production/purification collaborators

3 Cents/base

Cloning is long and boring, if I can, I try to have it done by others (company).

Gen9

Automated expression screening (1 person/1 week, 1152 cultures) Vincentelli et al. Methods 2011, Vincentelli. R and Romier. C, COSB, 2013, 23:1–9 N. Saez and Vincentelli. R MIMB, 2014

⇒  Possibility to test up to 1152 proteins in one culture condition or 1152 conditions on » one » protein (mutants, truncations…) on 1- 4 ml cultures. ⇒  ng-µg of proteins Day 1: Transform up to 1152 genes in one strain (12x96 transformations)

JOVE 2014

Open access video

Histag (GW, clone collection/synthetic genes)

TRX ≥ MBP > SUMO >NusA > GST > Histag (screen tags since 2002)

Day 2: Bacterial culture/induction from liquid transformation (12 DW96 or 48 DW24, 1152 cultures) Ros (DE3) pLys S, auto-induction, 25°C, 24 h

Day 3: Harvest culture Day 3/5 Lysis the 12xDW96 cultures and Purification : 12x Nickel 96 (30 min/ 96 purifications) Analysis on elution fractions on Capillary electrophoresis Caliper GX II (66 min/96 points) The P. Elmer Caliper GX II gives quantitative data (MW, concentration) (>100.000 points)

What we are using HTP expression screening for

1) 

Improve the success rate and protein soluble level

2)  3) 

Screen tens of constructs for a single protein or hundreds of protein in a single culture condition. Screen mutants, truncations libraries

4) 

Purify enough proteins for micro-assays (functional test, protein-DNA, protein-protein…)

5) 

Excellent Benchmarking tool. The culture and purification conditions (buffer, beads type and trademark…), are screened extensively and the same set up is used at scale up to have good reproductibility.

One person can work on hundreds of cultures easily, not so easy at the scale up culture/ purification/crystallization side.

Improving (co) –expression/purification in E.coli (I) Cell cultures

Expression vectors

Proteins Constructs Vectors Promoters Affinity tags Fusion proteins Tag/Fusion positions Number of Genes per vector Gene order on vector

Vector combinations Strains Helper plasmids Chaperones Growth media Inductor Inductor concentration Deepwells (24/96) Temperature Shaking orbital

Purification Vincentelli & Romier (2016) Advanced Technologies for Protein Complex Production 9

Affinity resin Magnetic / Chromatographic resin Resin volume Lysis (detergent) Purification buffer Centrifugation / Filtration Elution / Resin beads denaturation

Multi screen on a single culture Buffers Cell pellets (4 ml culture)

[Salt]

pH

Gradient / Sparse matrix

Lysis Affinity resin binding

Mid-size Culture (500 ml)

Purification

Vincentelli & Romier (2016)

Elution Results analysis

Advanced Technologies for Protein Complex Production 10

What we are using HTP expression screening for

1) 

Improve the success rate and protein soluble level by screening a large number of protein/constructs/culture conditions in parralel.

2) 

Screen tens of constructs for a single protein or hundreds of protein in a single culture condition (homologues, ORFeome, trunctation, mutation libraries...).

3)  Purify enough proteins for custom made micro-assays SELEX96 (Jolma et al, Cell, 2013), Pull-down/Hold up 384 (Vincentelli, Luck et al, Nature methods 2015) Initially purification on the tecan of ng, then µg, now mg of pure proteins from DW culture 4)  Excellent Benchmarking tool.

One person can work on hundreds of cultures easily, not so easy at the scale up culture/ purification/crystallization side so… The more we learn at analytical scale the better.

FROM VENOMS TO DRUGS

⇒  Identify new toxins from venomous animals ⇒  Try to Produce thousands of toxins by chemical synthesis or recombinant expression ⇒  Develop new (UHTP) protocols for correct folding of cystein rich proteins in E.coli ⇒  Identify high affinity (low nM) binders to cellular models for diseases to develop a new drug

venomics-project.eu 2012-2015

2012-2015

VENOMICS: a new paradigm high-throughput omics-based lead generation from venoms Marseille

Paris

Valencia, Spain

RNA cDNA

MS

Belgium

De novo MS Sequencing

Transcriptome cDNA libraries Sequences bank

Chemical Synthesis

E. coli Lisbon, Portugal

Peptide bank

Screening

Copenhagen, Danemark

Drug leads

SOURCING: 203 new species of venomous animals => Identified 25.000 new toxins

45 9 37

3

41 2 5

40

17

PRODUCTION (AFMB, WP Leader) Aim: Deliver pure oxydized toxins as powder or in water (endotoxin free) at 10 µM (250 ul) Toxins 35 to 120 AA, 1 to 9 SS Recombinant expression (1.8 M€): 35-120 AA, no PTM Protocol: Get pure oxydized toxins purified from E. coli

Production in E. coli: Back to our basics, benchmarked in HTP all known protocols and identified the best options for VENOMICS

Cell cultures

Expression vectors

Strains BL21(DE3) pLys S Ori(DE3) pLys S Shuffle. Co expression of Chaperones (PDI, Erv1p) Growth media Temperature/time of culture Culture scale

Codon optimisation/Gene design algorythm (3) cloning system (GW, LIC) Vector backbone Fusion proteins and vectors (12) His, HisMBP, HisGST, HisDsbA, His DsbC (Signal peptide)

Culture + Lysis + Nickel + TEV + Analysis Purification/cleavage Modified from

Vincentelli & Romier (2016) Advanced Technologies for Protein Complex Production

Lysis/purification buffer Affinity resin trademark Resin volumes Wash/elusion buffers and volumes Cleavage: 3 Proteases, ratio, time, buffer, cleavage sequence (TEV, no extra amino acid)

RTD phase: > 20.000 cultures/purifications/analysis in 2 years (2 FTE)

FINAL PRODUCTION

FINAL RECOMBINANT PRODUCTION PIPELINE (UHTP Production?) 3 FTE Recombinant expression: try to get as many oxidized toxins as possible from 5000 toxins (35-120, no PTM). Only 1 test/toxin possible (too many targets to try alternative protocol). Batch 1: 4000 to be produced in 6 months (January-June 2015) Delivered in July 2015 to the screening facility Batch 2: 1000 to be produced in 1 month (September 2015) Delivered end of September 2015 to the screening facility (project ended Oct 30, 2015)

NZYTech HTP Gene Synthesis platform

Developped an IP free vector family (LIC cloning) 2x96 synthesis/cloning per week This HTP platform is now opened to customers NZYTech © 2014

Results: Production of 4,992 synthetic genes using NZYTech HTP Gene synthesis platform

Cloning

3818

76.5 % correct by sequencing after checking only one colony 93 % correct by sequencing after checking only two colonies 809 365 Genes with 1 screened clone

NZYTech © 2014

Genes with 2 screened clones

Genes with 3 screened clones

Recombinant Strategy: purify oxidized toxins in E. coli from DsbC fusions

Cloning (NZYtech)

-  -  -  -  - 

Nozach et al, Microbial Cell Factories 2013, 12 (37) 12-16 Saez NJ et al, J Vis Exp. 2014 Saez NJ, Vincentelli R. Methods Mol Biol. 2014 Turchetto, Sequeira et al, Microbial Cell Factories in revision Turchetto, Sequeira et al, Microbial Cell Factories in revision

With the new protocol, during the RTD phase, the production of MT7, a difficult toxin to produce, was increased 550x Switching from the protocol published in 2013 to the new one (already 200 x for the cytoplasmic expression)

Culture conditions: BL21 (DE3) pLys S Nzytech auto-induction media 25°c 24h

Dsb C

-  Purification Nickel of His-DsbC-toxin -  -  -  -  - 

TEV cleavage C18 purification of Toxin LC-MS for QC/quantification Ajust concentration, aliquot for screening Freeze

Purification and QC: C18

Dsb C

Toxin Toxin

Toxin

PRODUCTION pipeline Spanning on 2 weeks for a single 96 plate (3 FTE)

Expression vectors: His-DsbC-Tev-Toxin

Day 1

Day 2

2x96 toxins/week

Transformation and preculture (Mon and Wed) PCR96

2 x 96 PCR 188 clones

Cultures in multiwells 2 x 96 toxins BL21 (DE3) pLys S, 25°C, Nzytech auto-induction media 96 DW24 1 toxin = 12 wells on a DW24 (24 ml) 2304 cultures Harvest and Freeze

2 x 96 toxins 96 DW24

Day 4

Purify (His) and Cleave DsbC overnight (Tev)

8 x Nickel 96 768 points

Day 5

Precipitate TEV with 0.1 % FA Purify toxin with C18 (SPE) in 50 % acetonitrile Evaporate overnight the acetonitrile to reach 250 ul

Day 3

Day 6

10 % of the prep is kept uncleaved and frozen in case of needs

Quantification and QC by LC-MS Adjust concentration, reformat and freeze for screeners

2304 cultures/week, EVERY WEEK during 7 months 60.000 cultures/10.000 purifications/5.000 LC-MS for the 5000 toxins

2 x C18_ 96 188 toxines 2 x LC-MS_ 96 188 toxines

RESULTS (52x96 toxins in production) Number of toxins Number of oxydized toxins

4992 2736

Average Fusion Yield (mg/L)

186

4.4 mg fusion purified before cleavage (1,6 g of TEV used during the production)

We purified a His-DsbC-(toxin) in 98 % of the cases but the final yield of toxin vs theory span from 10 to 100 % Is the Pipeline robust? 29/4992 (0,6 %) Did not transform or culture died during production 66/4992 (1,3 %) purification failed (well clogging, cross contamination). The culture were redone and all purified the second time ) 0%.

Some information we can extract

84 patterns (1 to 9 bridges), 17 with no toxin soluble (29 toxins,0.6 %) Yellow: produced Grey: not produced

38

38

40

55

39

76 77

62

62

60

45

24

35-40 40-60 60-80 80-100 > 100 AA AA AA AA AA (1304) (1185) (1086) (346) (82)

Toxin lengh

60 23

1 S=S (31)

36

41

61

64

59

38

2 S=S 3 S=S 4 S=S (95) (2084) (1078)

5 S=S (429)

56

54

43

46

6 S=S (257)

7 S=S (26)

Green: produced Red : not produced

100

100

8 S=S (2)

9 S=S (1)

Nav

NET

HTP Drug screening (July-October 2015) Protein Bank (SPPS+ E.coli): 3616 oxidized toxins

Cav Kv KCa

ASIC MSC

AR mAC HR NT-R (confidential targets)

-  10 screening assays on GPCR, ion channels, phenotypic and immunologic assays: Asthma, allergy, diabete, obesity, colon cancer, thyroid cancer, heart problems. CEA: in vitro assay on GPCR ZP: only in cellulo assays

30.000 tests analyzed, 400 hits including 20 in the nM range The bank will be screened by CEA/ZP on new assays when needed, potential hits will be validated further by ZP The bank will be on sale early 2017 for private companies outside of the consortium

-  - 

7 Scientific papers published in peer-reviewed journals (10 in preparation) > 150 publications in digital media/ newspapers (El Pais, Le Monde, Publico, Biofutur etc.) 11 TV (Euronews, France 5, Al Jazeera…) and radio programs (france info, RFI, RTBF…)

Wenkendavisen, Danemark

Le temps, Switzerland France 5 Al Jazeera

Euronews, January 19-24 (12 languages)

El Mundo, Spain El Pais

HTP production of the human PDZ repertoire, identification and ranking of the PDZ interacting with the Human papillomavirus E6 oncoprotein. Collaboration with: -  Dr G. Travé, UMR 7272, Strasbourg, France -  Dr J.P Borg, Dr J. Reboul, IPC, Marseille, France -  Dr N. Wolff and Dr Y. Jacob, Pasteur Institute, Paris, France See the seminar from G. Travé

Introduction – PDZ PDZ

SH3

GK

scaffolding role for the assembly of protein complexes

(picture adapted from Nourry et al. 2003, Science Signal.)

cell migration cell differentiation neurotransmission auditory sensory organ embryonic phototransduction development development

establishment and maintenance of cell polarity is a frequently observed function of proteins containing PDZ domains

Papillomavirus and cancer Dr G. Travé, UMR 7272, Strasbourg, France

mammals, birds & reptiles orogenital and cutaneous epithelia

80% 20%

low risk: warts and lesions

> 200 PV strains, > 100 HPVs Very small genome (8kb)

high risk:

skin cancers cervical cancer head & neck cancers

E6 & E7: the oncoproteins of HPV -> Cell proliferation priming

§  E6 and E7 bind / degrade MANY cellular proteins (>100) §  E6 and E7 hijack E3 ubiquitin ligases (E6AP, cullin 2...) §  E7 alters cell cycle checkpoints (Rb proteins, cyclins…) §  E6 alters apoptosis control (p53, Bak…) §  E6 alters cell adhesion pathways (PDZ-containing proteins)

=> domain-mo+f interac+ons

E6 hijacks domain-mo+f networks

E6 C-terminus hooks PDZ domains (cell communication, adhesion, signalling)

(CELLULAR) Zn2+

(VIRAL)

PDZ 1

Motif Mimicry

Zn2+ 151

1

E6N

E6C

PDZ 2 PDZ 3

[xTxL/V-COOH]

PDZ 4

PDZ 5

Iden+fica+on and ranking of E6-binding PDZ domains PDZ 1

Zn2+

PDZ 2

Zn2+

E6 C-term

151

1

[xTxL/V-COOH]

PDZ i

HPV E6 All mucosal high-risk HPV E6s contain a PDZ Binding Motif Human « PDZome »: 266 PDZ domains found in 150 proteins

PDZ i+1

PDZ n

-Can we experimentally iden+fy all E6-binding PDZ domains -Can we hierarchize their affini+es for HPV E6 1) We need to produce as many human PDZ domains as possible 2) We need an HTP assay for systema+c PDZ-pep+de affinity determina+on

PDZ

Human PDZome production (266 PDZ) - 

Cloning (his-MBP) of the human PDZome (266 PDZ) from a Gateway clone collection (PDZome, Zimmerman et al PLoS One. 2013;8(2), -  Production (ZYP5052) on 4 ml cultures in triplicates (798 cultures in one go). =>PDZome V1 - 

After lysis, the PDZ crude supernatants are loaded on Caliper, the concentration of the soluble PDZ are calculated, adjusted to 4 µM (tecan) and frozen in 96 (12 ml culture/30 copies of the PDZome).

223/266 (84%) PDZ soluble above 4 µM in standard culture conditions 241/266 (91%) soluble above 4 µM after changing culture conditions/purifications. To try improving the PDZome we made a new clone collection =>PDZome V2 -redesign of PDZ Sequences (Luck et al 2012), optimized for E. coli

PASTEUR PDZome (N. Wolff) PDZ V2 : 264/266 soluble, 2 PDZ missing in V2 were fine in V1: 266/266 soluble.

6 ml of cultures : 80 copies of PDZome V2, 4-6x more proteins with PDZome V2 than PDZome V1. On average the PDZome V2 is >170 mg/L of cultures.

HTP In vitro Protein-protein interaction studies Automated pull down and Hold up assays (HU) The Hold up is designed to determine weak interactions (µM) at equilibrium and affinity determination (Charbonnier, and Trave, G. (2006), Protein Expr Purif 50, 89-101)

HTP Hold up (or pull down) (2007) + caliper detection (2010) Filter plates

8x 96 binary interactions/day Validated with Histag, GST, Streptavidin fusions on pure proteins or crude extract for both proteins. (10 µl of resin/point)

No robot, no problem, still easy to reach few hundred points/week by hand Since 2012 Validated on 384 millipore filter plate (2,5 µl resin, 4µl protein) Max: 6x 384 (2304 points) binary interactions/day We are now limited by the P. Elmer caliper GX II (“only”2x384 points/day) in analysis

Human PDZome vs HPV16/HPV18 The Hold up is quantative and allow to rank affinities Charbonnier, and Trave, G. (2006), Protein Expr Purif 50, 89-101

Interaction 266 hPDZ vs 6 peptides: 1 day on robot, 3 days on caliper

12 known => 11 confirmed by HU + 38 new ones

We estimate all K values. The values for the best binders range from 1 uM to 200 uM. D High-throughput determination of domain-ligand affinities by automated holdup assay allows quantification of linear motif specificities. R. Vincentelli*, R. Vincentelli*, K. Luck*, J. Poirson, J. Polanowska, J. Abdat, M. Blémont, J. Turchetto, F. Iv, K. Ricquier, M.L Straub, A. Forster, P. Cassonnet, J.P Borg, Y. Jacob, M. Masson, Y. Nominé, J. Reboul, N. Wolff, S. Charbonnier, G.Travé. Nature Methods, 2015 Aug;12(8):787-93

ValidaDon by orthogonal methods V1 Y2H and mammalian GPCA

- an European PhD Training Network

Exploring PDZ domain proteins as novel therapeutic intervention points

12 partners (3 industrial) 14 PhD fellows 6 countries €3.75 mill. 4 years (2016-2020) www.pdznet.eu

EMBO practical courses 2005-2017 (every 2 years) Marseille: 2005, 2011, 2015

Oxford (OPPF): 2007, 2009, 2013 (SGC): Jult 6-14th, 2017 (application opens soon)

Marseille: Nov 2013 FRISBI course (French INSTRUCT network) Protein Production for Protein Complexes

The Protein Expression team Fundings: - ANR PFTV 2008-2011 -  ANR IBISA 2009-2011 Customers and collaborators (academic and private companies)

Former Scientific Coordinator Dr C. Cambillau

Scientific coordinator Dr A. Roussel

Manager Dr R. Vincentelli

ANR HPV16/PDZ 2008-2010 IGBMC, Strasbourg, France Dr G. Travé, Dr S. Charbonnier, Dr K. Luck Pasteur, Paris, France Dr N. Wolff AFMB Julie Abdat, Marilyne Blémont, Jéremy Turchetto PDZome V1 PDZome V2

CRCM, Marseille, Dr J.P Borg NZYtech, Lisbon, Portugal, C. Fontes

VENOMICS H. Darbon N. Saez

2002

2009 J. Turchetto

“Bob and Candy”

F. Peysson

L. Ramond