Differential transcriptional responses to Ebola and Marburg virus infection in cells from bats and humans

Supplemental Material

Martin Hölzer, Verena Krähling, Fabian Amman, Emanuel Barth, Stephan H. Bernhart, Victor Carmelo, Maximilian Collatz, Gero Doose, Florian Eggenhofer, Jan Ewald, Jörg Fallmann, Lasse Moritz Feldhahn, Markus Fricke, Juliane Gebauer, Andreas J. Gruber, Franziska Hufsky, Henrike Indrischek, Sabina Kanton, Jörg Linde, Nelly Mostajo Berrospi, Roman Ochsenreiter, Konstantin Riege, Lorena Rivarola-Duarte, Abdullah H. Sahyoun, Sita J. Saunders, Stefan E. Seemann, Andrea Tanzer, Bertram Vogel, Stefanie Wehner, Michael T. Wolfinger, Rolf Backofen, Jan Gorodkin, Ivo Grosse, Ivo Hofacker, Steve Hoffmann, Christoph Kaleta, Peter F. Stadler, Stephan Becker, Manja Marz

Contents

• ES 1: Data
• ES 1A: Samples and Libraries
• ES 1B: Transcriptome assembly: Rousettus aegyptiacus
• ES 2: Differential gene expression (DGE)
• ES 2A: Hierarchical clustering
• ES 2B: Overview
• ES 2C: Timepoints
• ES 2D: Contrasts
• ES 2E: Human vs Bat cells
• ES 2F: Human vs Filovirus with replicates
• ES 2G: Human vs Bat cells with replicates
• ES 2H: De novo genes
• ES 2I: ncRNAs
• ES 3: Genes viewer
• ES 3A: protein-coding genes and ncRNAs
• ES 3B: de novo genes
• ES 4: Common genes during filovirus infection
• ES 4A: Summary (PDF)
• ES 4B: Scatterplots
• ES 4C: Group plots (time series)
• ES 5: MARA analysis of transcription factors and binding sites
• ES 5A: EBOV and MARV
• ES 5B: EBOV only
• ES 5C: MARV only
• ES 5D: Summary (PDF)
• ES 6: Pathways
• ES 7: qRT-PCR
• ES 7A: Primer design
• ES 7B: qRT-PCR results
• ES 7C: Melt curves
• ES 7D: Gels
• ES 7E: IFA of second experiment

Download Supplemental Material: pdf (~1.6MB)

ES 1A: Samples and Libraries

All reads were preprocessed for quality, 5'-bias and poly-A tails. Preprocessed HiSeq libraries for HuH7 and R06E-J samples were pooled for assembly.

Additionally, all nine R06E-J samples were pooled and sequenced on a single MiSeq lane with a length of 300 bp and in paired-end mode. Read coverage was calculated on an expected eukaryotic transcriptome size of 75 Mbp.

Sample	Read Numbers		Coverage		FastQC read statistics
	raw	trimmed	raw	trimmed	raw	trimmed
HuH7: HiSeq Samples
HuH7-MOCK-3h	40,529,770	38,408,644	54X	51X	forward,reverse	forward,reverse
HuH7-MOCK-7h	38,008,320	35,958,522	51X	48X	forward,reverse	forward,reverse
HuH7-MOCK-23h	39,006,442	36,866,384	52X	49X	forward,reverse	forward,reverse
HuH7-EBOV-3h	34,429,922	32,777,274	46X	44X	forward,reverse	forward,reverse
HuH7-EBOV-7h	49,982,536	46,619,978	67X	62X	forward,reverse	forward,reverse
HuH7-EBOV-23h	53,034,360	50,112,880	71X	67X	forward,reverse	forward,reverse
HuH7-MARV-3h	44,197,622	41,778,332	59X	56X	forward,reverse	forward,reverse
HuH7-MARV-7h	48,264,544	45,678,616	64X	61X	forward,reverse	forward,reverse
HuH7-MARV-23h	36,284,714	34,261,494	48X	46X	forward,reverse	forward,reverse
HuH7-Pooled	382,332,302	362,462,124	510X	483X	forward,reverse	forward,reverse
R06E-J: HiSeq Samples
R06E-J-MOCK-3h	50,396,276	47,815,138	67X	64X	forward,reverse	forward,reverse
R06E-J-MOCK-7h	44,016,800	41,351,362	59X	55X	forward,reverse	forward,reverse
R06E-J-MOCK-23h	48,464,382	45,474,028	65X	61X	forward,reverse	forward,reverse
R06E-J-EBOV-3h	41,450,606	39,376,162	55X	53X	forward,reverse	forward,reverse
R06E-J-EBOV-7h	38,362,126	36,005,676	51X	48X	forward,reverse	forward,reverse
R06E-J-EBOV-23h	39,292,722	37,350,338	52X	50X	forward,reverse	forward,reverse
R06E-J-MARV-3h	37,490,486	35,538,058	50X	47X	forward,reverse	forward,reverse
R06E-J-MARV-7h	48,645,544	45,862,636	65X	61X	forward,reverse	forward,reverse
R06E-J-MARV-23h	45,569,356	43,291,468	61X	58X	forward,reverse	forward,reverse
R06E-J-Pooled	393,021,564	372,082,040	524X	496X	forward,reverse	forward,reverse
R06E-J: MiSeq Sample
R06E-J-MiSeq	38,163,200	38,028,488	153X	152X	forward,reverse	forward,reverse

ES 1B: De novo transcriptome assembly of Rousettus aegyptiacus

Assembly	Contigs	Contigs >1000 bp	max. Contig (bp)	Parameter	Download (*.tar.gz)
Velvet/Oases*	370,200	180,458	36,073	k = 45,55,65,75; multiple kmer-assemblies of Velvet merged by Oases like described in manual	link (95 Mb)
SoapDenovo-Trans*	699,418	147,144	34,307	k = 25,35,45,55,65,75; multiple kmer-assemblies merged by cd-hit-est (-c 0.95)	link (189 Mb)
Trinity*	484,826	188,534	30,056	k = 25 (fix)	link (172 Mb)
ABySS/TransABySS*	790,204	169,324	21,424	k = 25,35,45,55,65,75; multiple kmer-assemblies of ABySS merged by TransABySS pipeline	link (192 Mb)
Mira**	162,861	21,987	14,119	overlap graph approach; job = est,denovo,accurate	link (33 Mb)
Clustered (cd-hit-est)	977,787	277,595	36,073	-c 0.95	link (290 Mb)

* 9 different samples were sequenced individually on a Illumina HiSeq with 100bp, paired end; after sequencing pooled for assembly

** all 9 samples were pooled and normalized before library preparation, then sequenced on a Illumina MiSeq, 300bp, paired end

QUAST: full report statistics

ES 2A: Hierarchical clustering of log2 gene-expression fold-changes upon infection of human cells with EBOV or MARV

Hierarchical clustering of log2 gene-expression fold-changes upon infection with EBOV or MARV, respectively. Deregulated genes (padj < 0.1) with an absolute log2 expression fold-change greater than five in at least one of the contrasts (infection versus Mock control at a specific time point post infection, p.i.) have been considered for hierarchical clustering of log2 gene expression fold changes.

• Download: pdf
• Download: svg

ES 2B: Differential gene expression: Overview

Pairwise Comparisons

Overview of different comparisons within human (bat) samples and between the two species. Click on button to see all comparisons and links to the corresponding tables with all differential expressed genes with a padj < 0.1. Same colors indicate that the results are listed in the same table.

	MOCK-3h	MOCK-7h	MOCK-23h	EBOV-3h	EBOV-7h	EBOV-23h	MARV-3h	MARV-7h	MARV-23h
MOCK-3h
MOCK-7h
MOCK-23h
EBOV-3h
EBOV-7h
EBOV-23h
MARV-3h
MARV-7h
MARV-23h

Comparisons including Replicates

Overview of different comparisons between Mock and filovirus (EBOV, MARV) samples using replicates. Click on button to see all comparisons and links to the corresponding tables with all differential expressed genes filtered by a padj <= 0.05. Same colors indicate that the results are listed in the same table.

• MOCK = MOCK-3h, MOCK-7h and MOCK-23h as replicates
• FILO-xh = EBOV-xh, MARV-xh as replicates

	MOCK	FILO-3h	FILO-7h	FILO-23h
MOCK
FILO-3h
FILO-7h
FILO-23h

ES 2C: Comparisons between time points 3h, 7h and 23h of Mock, EBOV and MARV samples within human and bat

Differential gene expression of HuH7/R06E-J MOCK, EBOV and MARV samples compared at 3h, 7h and 23h, respectively. These tables show the differential expression within MOCK, EBOV and MARV and between the different time points, separate for human and bat. Tables are sorted by fold change.

• HuH7 Mock|EBOV|MARV
• Go to table
• Download CSV
• R06E-J Mock|EBOV|MARV
• Go to table
• Download CSV

ES 2D: Differential gene expression: Comparisons between Mock, EBOV and MARV samples

All tables are sorted by fold change.

Comparison	Table link	Download table	Description
HuH7 Mock/EBOV @3h,7h,23h	Go	click	Differential gene expression of HuH7 MOCK and EBOV samples compared at 3h, 7h and 23h post infection.
HuH7 Mock/MARV @3h,7h,23h	Go	click	Differential gene expression of HuH7 MOCK and MARV samples compared at 3h, 7h and 23h post infection.
HuH7 EBOV/MARV @3h,7h,23h	Go	click	Differential gene expression of HuH7 EBOV and MARV samples compared at 3h, 7h and 23h post infection.
R06E-J Mock/EBOV @3h,7h,23h	Go	click	Differential gene expression of R06E-J MOCK and EBOV samples compared at 3h, 7h and 23h post infection.
R06E-J Mock/MARV @3h,7h,23h	Go	click	Differential gene expression of R06E-J MOCK and MARV samples compared at 3h, 7h and 23h post infection.
R06E-J EBOV/MARV @3h,7h,23h	Go	click	Differential gene expression of R06E-J EBOV and MARV samples compared at 3h, 7h and 23h post infection.

ES 2E: Comparisons between human and bat

We compared differential expression between human and bat EBOV and MARV infected samples at 3h, 7h and 23h poi. Additionally, we used EBOV and MARV samples at same time points as replicates with DESeq to compare the characteristics of filovirus infection between human and bat.

• HuH7 EBOV 3h,7h,23h compared with R06E-J EBOV 3h,7h,23h
• Go to table
• Download CSV
• HuH7 MARV 3h,7h,23h compared with R06E-J MARV 3h,7h,23h
• Go to table
• Download CSV
• HuH7 EBOV+MARV 3h,7h,23h compared with R06E-J EBOV+MARV 3h,7h,23h
• Go to table
• Download CSV

ES 2F: Human MOCK samples compared to EBOV+MARV cells using replicates

Differential gene expression of HuH7 MOCK samples (3h, 7h and 23h as replicates) compared to EBOV+MARV samples (as replicates, called FILO) at 23h. For 3h and 7h no differential expressed genes were found with a padj < 0.05.

• HuH7 MOCK_3h,7h,23h vs. HuH7 FILO_{EBOV-23h,MARV-23h}
• Go to table
• Download CSV

ES 2G: Human vs Bat samples including replicates

Differential gene expression of HuH7 MOCK/EBOV/MARV samples compared to R06E-J MOCK/EBOV/MARV samples at 3h, 7h and 23h (filtered by a padj < 0.05). Within this experiment we compared:

1. HuH7 MOCK_3h,7h,23h vs R06E-J MOCK_3h,7h,23h
• Go to table
• Download CSV
2. HuH7 FILO_{EBOV-3h,MARV-3h} vs R06E-J FILO_{EBOV-3h,MARV-3h}
• Go to table
• Download CSV
3. HuH7 FILO_{EBOV-7h,MARV-3h} vs R06E-J FILO_{EBOV-7h,MARV-3h}
• Go to table
• Download CSV
4. HuH7 FILO_{EBOV-23h,MARV-3h} vs R06E-J FILO_{EBOV-23h,MARV-3h}
• Go to table
• Download CSV

and additionally filterd results of (2), (3) and (4) by removing genes which were also found as differential expressed between MOCK samples (1):

• MOCK Filtered: HuH7 FILO_{EBOV-3h,MARV-3h} vs R06E-J FILO_{EBOV-3h,MARV-3h}
• Go to table
• Download CSV
• MOCK Filtered: HuH7 FILO_{EBOV-7h,MARV-3h} vs R06E-J FILO_{EBOV-7h,MARV-3h}
• Go to table
• Download CSV
• MOCK Filtered: HuH7 FILO_{EBOV-23h,MARV-3h} vs R06E-J FILO_{EBOV-23h,MARV-3h}
• Go to table
• Download CSV

ES 2H: Comparison of de novo predicted gene loci of human and bat

We used the first step of the Cufflinks pipeline to detect possible gene loci de novo based on Tophat mapped read data. We further compared the differential expression on each gene loci between different time points and within one condition as well as at same time points and different conditions. All results for human and bat (norm. read counts, log2 fold changes, padj) are listed in the following tables, respectively.

• Human
• Go to table
• Download CSV
• Bat
• Go to table
• Download CSV

ES 2I: Comparison of ncRNA genes within human and bat

We compared the differential expression of each ncRNA between different time points and within one condition as well as at same time points and different conditions. All results for human and bat (norm. read counts, log2 fold changes, padj) are listed in the following tables, respectively.

• Human
• Go to table
• Download CSV
• Bat
• Go to table
• Download CSV

ES 4: Common genes during filovirus infection

ES 4A: Summary download (pdf)

• Human 3h
• Human 7h
• Human 23h
• Bat 3h
• Bat 7h
• Bat 23h

ES 4B: Scatterplots

To compare the differential expression of Mock/EBOV and Mock/MARV in human and bat cells, log2 fold changes (FC) as computed by DEseq were visualized using scatter plots in R (x-value: FC of Mock/EBOV; y-value: FC of Mock/MARV). The scatterplots were overlaid with contour plots for a two-dimensional kernel estimate (kde2d; MASS package) using the default parameters. Outliers were labeled with the respective gene names.

	3h	7h	23h
Human
Bat

ES 4C: Group plots/Timeseries

To identify co-regulated genes over the course of the three timepoints (3h, 7h, 23h) the differential expression of Mock/EBOV (x-value) and Mock/MARV (y-value) was visualized for all genes using lines. The expression changes between the three time points define an angle that is subsequently used to group genes that have approximately the same or approximately inverse expression time courses. Note, that the angle between the three time points is rotation invariant.

	3h	7h	23h
Human
Bat

ES 6: Pathways

How to read pathway figures

Here, we describe how to read the gene-regulation boxes next to each gene, to fastly follow the pathways and their regulations over all time points, infections and hosts. Shown are the most significant genes according to expression level differences. To give an example how to read the boxes displayed here and in the different pathways in the electronical supplement, we use DUSP8 for explanation:

The upper half (pink) displays human samples, the lower half (blue) refers to bat samples. DUSP8 is in human Mock cells is slightly down-regulated from 3 to 7h, furthermore slightly down-regulated from 7 to 23h. The minimum of transcripts for all three timepoints in human Mock cells is read_max=31. In EBOV-infected human cells we observe from 3-7h a slight down-regulation, although the manual profile impression leads a slight up-regulation. From 7 to 23h we see a dramatic up-regulation of 25X leading to a read_max of 626. For MARV-infected human cells we see from 3-7h a slight down-regulation similar to EBOV-infected human cells from 3-7h, however this mRNA level stays from 7 to 23h constant. In bats we observe in Mock from 7 to 23h a slight down-regulation, similar to EBOV-infected bat cells from 3h to 7h. In a later stage, we observe a slight up-regulation, just as during the complete MARV-infected bat cells. The regulation level is only changing slightly as depicted in the manual search.

Furthermore, we framed each gene red, when being differentially expressed in human between the different time points. We added a green frame, when a difference between EBOV and MARV was observed in the general expression and we used blue to mark differences between human and bat transcriptome levels. The line width of the frame corresponds to the significance of the expressional change.

NA - not assigned; no bat homologous gene was detected. When the expression level changes by >15% an arrow indicates the regulation; when the expression level changes by more than 100% (2-fold transcription) a number beside the arrow indicates the fold change.

Investigated Pathways

S6.1: The Filovirus Infection Network S6.2: MAPKs S6.3: Activation of MAPKs S6.4: MAPK (p38) S6.5: MAPK and cell migration (JNK) S6.6: MAPK and cell migration (ERK) S6.7: MAPK and cell migration (P38) S6.8: Pathway induced by RNA-virus S6.9: General immune response S6.10: NFκb activation by virus S6.11: Viral interactions with nucleocytoplasmic transport pathways S6.12: Unfolded protein response within the endoplasmatic reticulum S6.13: Cell cycle during Epstein-Barr virus infection S6.14: Herpes simplex infection S6.15: MyD88 S6.16: CREB initiation

S6.17: NFκB S6.18: Focal adhesion S6.19: Actin pathway S6.20: GADD34 S6.21: DUSP S6.22: JAK/STAT S6.23: Insulin S6.24: ARL13B S6.25: Endocytosis S6.26: mTor S6.27: PI3K Akt S6.28: PI3K NFκB S6.29: TGF-B interactions S6.30: RIP3/RIP1-mediated necrosis S6.31: Cell cycle, Peroxisome, Steroid biosynthesis, complement and coagulation cascades S6.32: Endocytosis S6.33: Apoptosis

Overview of all kingpins and related interactions svg pdf and description
Mitogen-activated protein kinase (MAPK) svg pdf and description		Activation of MAPKs svg pdf and description		MAPK (p38) svg pdf and description
MAPK and cell migration (JNK) svg pdf and description		MAPK and cell migration (ERK) svg pdf and description		MAPK and cell migration (P38) svg pdf and description
Pathway induced by RNA-virus svg pdf and description		Pathway for general immune response svg pdf and description		NFκB activation by viruses svg pdf and description
Viral interactions with nucleocytoplasmic transport pathways svg pdf and description		Unfolded protein response within the endoplasmatic reticulum svg pdf and description		Cell cycle during Epstein-Barr virus infection svg pdf and description
Herpes simplex infection svg pdf and description		MyD88 svg pdf and description		CREB initiation svg pdf and description
NFκB svg pdf and description		Focal adhesion svg pdf and description		Actin pathway svg pdf and description
GADD34 svg pdf and description		DUSP svg pdf and description		JAK/STAT svg pdf and description
Insulin svg pdf and description		ARL13B svg pdf and description		Endocytosis svg pdf and description
mTor svg pdf and description		PI3K Akt svg pdf and description		PI3K NFκB svg pdf and description
TGF-B interactions svg pdf and description		RIP3/RIP1-mediated necrosis svg pdf and description		Cell cycle, Peroxisome, Steroid biosynthesis, complement and coagulation cascades svg pdf and description
Endocytosis svg pdf and description		Apoptosis svg pdf and description

ES 7: qRT-PCR

qRT-PCR were performed for the human and bat homologous genes HAVCR1, NPC1 and TLR3. Please note that for the qRT-PCR new samples from an similar infection experiment of the same cell lines were used. We noticed that the infection rate was lower as in the cells used within the transcriptome study.

ES 7A: Primer Design

• Primer sequences: FASTA
• CDS sequences (human): FASTA
• CDS sequences (bat): FASTA

ES 7B: qRT-PCR values and 18S normalized mRNA abundances

Download EXCEL

ES 7C: Melting curves

• H. sapines: MH3 = Mock 3h | MH23 = Mock 23h | ZH3 = ZEBOV 3h | ZH23 = ZEBOV 23h
• R. aegyptiacus: MR3 = Mock 3h | MR23 = Mock 23h | ZR3 = ZEBOV 3h | ZR23 = ZEBOV 23h

NPC1	HAVCR1	TLR3

ES 7D: Agarose gel

2% agarose gel in TAE was run in 120V for 35 minutes. (-): control; water

On first gel, primer for bat HAVCR1 did not amplify. HAVCR1 was rerun on a second gel.

ES 7E: IFA

We repeated the infection of HuH7 and R06E-J cells and isolated RNA for qRT-PCR analyses. An IFA of these cells revealed that the infection rates were lower than in the previous experiment used for RNA-Seq.