Modelling the Acinetobacter baumannii colistin antibiotic stress-response at the metabolic level Luana Presta , Emanuele Bosi , Leila Mansouri , Lenie Dijkshoorn , Renato Fani and Marco Fondi a
a
a
a
b
a
a
Department of Biology, University of Florence, Florence, Italy; b Department of Infectious Diseases, Leiden University Medical Center, Leiden, The Netherlands
INTRODUCTION Acinetobacter baumannii is a clinical threat to human health, causing major infection outbreaks worldwide. One of the last-line treatments for MDR A. baumannii is represented by colistin, a positively charged molecule that, by interacting with the lipid A moiety of lipopolysaccharide (LPS), causes disorganization of the outer membrane. Unfortunately, colistin resistance in A. baumannii has been reported, thus highlighting the urgency to find new molecules facing this threat.
GEM and Flux Balance Analysis
AIMS AND METHODS
1. Draft reconstruction
Since new drugs against gram-negative bacteria seems not to be forthcoming, there is an urgent need for nove potential therapeutic targets. Genome-scale metabolic network (GEM) reconstruction and its conversion to a mathematical framework allows to examine the connection between phenotype and genotype and to drive biological discoveries. In particular, constraint-based tools (such as Flux Balance Analysis, FBA) enable to estimate metabolites’ flow through a metabolic network and to compute cellular phenotypes for various growth conditions, such as antibiotic treatment and genome-scale gene deletion.
Data assembly and dissemination 95| Print Matlab model content. 96| Add gap information to the reconstruction output.
1| Obtain genome annotation. 2| Identify candidate metabolic functions. 3| Obtain candidate metabolic reactions. 4| Assemble draft reconstruction. 5| Collect experimental data.
TARGET
GENE
MATHEMATICAL MODELING v3
Constraints 1) Sv = 0 2) a i < v i < b i
v3
Unconstrained solution space
v2
Allowable solution space
Whole genome sequence
• Previous models • Primary literature
v1 Optimal solution
Genome-scale metabolic model
Metagenomic sequence reads
• Public databases
3. Conversion of reconstruction into computable format
1
38| Initialize the COBRA toolbox. 39| Load reconstruction into Matlab. 40| Verify S matrix. 41| Set objective function. 42| Set simulation constraints.
Gene-protein-reaction associations 4
A
2
Thiele, Ines, and Bernhard Ø. Palsson. "A protocol for generating a high-quality genome-scale metabolic reconstruction." Nature protocols 5.1 (2010): 93-121.
v3
v1
v1
v2
Optimization maximize Z
CLINICAL-TRIALS
DRUG
4. Network evaluation 43−44| Test if network is mass-and charge balanced. 45| Identify metabolic dead-ends. 46−48| Perform gap analysis. 49| Add missing exchange reactions to model. 50| Set exchange constraints for a simulation condition. 51−58| Test for stoichiometrically balanced cycles. 59| Re-compute gap list. 60−65| Test if biomass precursors can be produced in standard medium. 66| Test if biomass precursors can be produced in other growth media. 67−75| Test if the model can produce known secretion products. 76−78| Check for blocked reactions. 79−80| Compute single gene deletion phenotypes. 81−82| Test for known incapabilites of the organism. 83| Compare predicted physiological properties with known properties. 84−87| Test if the model can grow fast enough. 88−94| Test if the model grows too fast.
2. Refinement of reconstruction 6| Determine and verify substrate and cofactor usage. 7| Obtain neutral formula for each metabolite. 8| Determine the charged formula. 9| Calculate reaction stoichiometry. 10| Determine reaction directionality. 11| Add information for gene and reaction localization. 12| Add subsystems information. 13| Verify gene−protein-reaction association. 14| Add metabolite identifier. 15| Determine and add confidence score. 16| Add references and notes. 17| Flag information from other organisms. 18| Repeat Steps 6 to 17 for all genes. 19| Add spontaneous reactions to the reconstruction. 20| Add extracellular and periplasmic transport reactions. 21| Add exchange reactions. 22| Add intracellular transport reactions. 23| Draw metabolic map (optional). 24−32| Determine biomass composition. 33| Add biomass reaction. 34| Add ATP-maintenance reaction (ATPM). 35| Add demand reactions. 36| Add sink reactions. 37| Determine growth medium requirements.
PATIENT-CARE
PRECLINICAL
Karlsson, Fredrik H., et al. "Prospects for systems biology and modeling of the gut microbiome." Trends in biotechnology 29.6 (2011): 251-258.
B
ATP ADP
E1
P1
E2
P2
E3
P3
C
NAD + NADH
E4 D
0
3
P4
1 -1 0 1
P5
Reaction Enzyme Protein
v2
S
Stochiometric matrix
0
Gene
0 0
1
1
0
0
0
0
0
1
-1
0
0
0
0
0
1
0
0
1
0
-1
-1
-1
0
0
0
0
-1
0
1
0
-1
0
1
0
0
1
0
0
1
0
MODEL VALIDATION AND RESULTS
3
34
MOMA FBA
0
3) Predicted EGs are consistent with available experimental datasets Essential genes (EGs) found with in silico simulation have been compared to those previously found with both wet-lab and in silico techniques, for A. baumannii ATCC 19606, A. baumannii ATCC 17978 and A. baumannii AYE.
68
iLP1043 Rich medium
57
iLP1043
21 36 225
ATCC19606
117
600 Genes
800
1000
0
ATCC17978
0
iLP1043
400
iLP1043
AbyMBEL891
224
8 6
46
MOMA FBA
200
400 600 Genes
1.0
1.0
0.5
0.5
0.0
−0.5
−1.0
−1.0
116 −0.5
0.0
0.5
−1.0
1.0
−0.5
5) Colistin exposure changes predicted metabolic fluxes in central A. baumannii pathways 60% of reactions appear to be affected by colistin treatment, since fluxes variation between the control and the treated model has emerged. 60’
15’
60’
15’
Steady
60’
15’
Increasing
60’
15’
Decreasing
60’
15’
On
60’ Off
Vitamin B6 metabolism Valine, leucine and isoleucine biosynthesis Urea cycle and metabolism of amino groups Taurine and hypotaurine metabolism Synthesis and degradation of ketone bodies Starch and sucrose metabolism Riboflavin metabolism Pyruvate metabolism Pyrimidine metabolism Purine metabolism Propanoate metabolism Porphyrin and chlorophyll metabolism Phenylalanine tyrosine and tryptophan biosynthesis Peptidoglycan biosynthesis Pentose phosphate pathway Pentose and glucuronate interconversions Pantothenate and CoA biosynthesis One carbon pool by folate Nucleotide sugars metabolism Nitrogen metabolism Nicotinate and nicotinamide metabolism Lysine biosynthesis Glyoxylate and dicarboxylate metabolism Glycolysis Glycine serine and threonine metabolism Glycerophospholipid metabolism Glycerolipid metabolism Glutathione metabolism Glutamate metabolism Fructose and mannose metabolism Folate biosynthesis Fatty acid metabolism Fatty acid biosynthesis D−Glutamine and D−glutamate metabolism D−Alanine metabolism Citrate cycle (TCA cycle) Butanoate metabolism Benzoate degradation via hydroxylation Benzoate degradation via CoA ligation Arginine and proline metabolism Aminosugars metabolism Alanine and aspartate metabolism
iLP1043 Untreated15’
60 minutes
8
7,87% 40,45% 31,24% 4,49% 9,66% 6,29%
0
5
10 15 0
5
10 15
0
5
10 15 0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15 0
5
10 15
0
5
10 15 0
5
10 15
0
5
10 15 0
5
0.5
1.0
6) Genome-scale gene deletion predicts new potential drug targets in colistin resistant A. baumannii ATCC 19606 Colistin resistant (LPS deficient strain) phenotype has been simultaed. TRanscriptomic data in such conditions have been merged to the model, this leading to identification of new potential drug-targets
15 minutes
8,04% 41,52% 27,17% 4,78% 11,96% 6,52%
0.0
Treated 60’ (x 103)
Treated 15’ (x 103)
Changing direction
1000
0.0
−0.5
−1.0
15’
800
4) Antibiotic treatment defines condition-specific models Availaible transcriptiomic data on A. baumannii ATCC 19606 (wild type and LPS-) response to colistin treatment (2) have been merged in to the metabolic model allowing to observe that changes in gene expression are likely to influence the activity rate of the corresponding cellular metabolic reactions.
Untreated 15’ (x 103)
iLP1043 Simmons medium
200
0.0 0.2 0.4 0.6 0.8 1.0
5
Untreated 60’ (x 103)
25
Simmons medium
GRratio
Growth
No growth
No growth
Model’s predictions
Growth
Rich medium 0.0 0.2 0.4 0.6 0.8 1.0
The growth rates were firstly in silico assessed in minimum Simmons medium by iteratively probing each C-source previously tested in PM plates, under aerobic conditions. Following this procedure, we finally reached an overall agreement (about 88%) with experimental data.
Experimental data
2) Constraint-based modelling identifies metabolic EGs Essential genes (EGs) (those genes whose deletion is predicted to impair cellular growth) have been identified by performing in silico gene deletion, on genome-scale, in different growth conditions.
GRratio
1) Genome-scale A. baumannii ATCC 19606 model (iLP1043) is consistent with large scale phenotypic data Consistency of model's predictive capability was verified by rigorously comparing FBA outcomes with experimentally determined auxotrophies data by Phenotype Microarray (PM) (1).
iLP1043 Treated 15’
57
8
iLP1043 Untreated 60’
9
iLP1043 Treated 60’
57
2
Putative targets for antibiotic treatments
10 15
N. of reactions
CONCLUSIONS A comprhensive and accurate genome scale metabolic model of A. baumannii ATCC 19606 metabolic model has been reconstructed. By integrating gene expression data with constraint-based modelling we described the metabolic reprogramming occurring after colistin-exposure in A. baumannii and the changes in the pattern of gene essentiality during such stress condition. Some of these genes, although not yet experimentally endorsed, might represent primary targets for future research on the treatment of both the wild type and LPS-mutant (i.e. colistin resistant) strains. Our results have practical implications for the identification of new therapeutics as the identified essential genes can be used in drug-design pipelines.
REFERENCES 1) Peleg, A. Y. et al. The success of Acinetobacter species; genetic, metabolic and virulence attributes. PLoS One 7, e46984 (2012). 2) Henry, R. et al. The transcriptomic response of Acinetobacter baumannii to colistin and doripenem alone and in combination in an in vitro pharmacokinetics/pharmacodynamics model. J. Antimicrob. Chemother. 70, 1303–1313 (2014).
[email protected] Scientific Reports 7, Article number: 3706 (2017) doi:10.1038/s41598-017-03416-2