Microbial Ecology

TWiM #274: Bacterial Endosymbionts Block Giant Viruses

Podcast and Annotation Information
  • Annotation by Kylie Lynch, Zoe E. Johnson, Grace Helle, Michaela Gazdik Stofer,  and Katriana A. Popichak.
  • Podcast audio by TWiM: Listen to TWiM #274 Podcast
  • Podcast transcript by Kylie Lynch, Zoe E. Johnson, Katriana A. Popichak, Aidan Sasaoka, and Laurel Thompson: Access Podcast Transcripts
  • Papers Discussed:
    • Cheng AG, Ho PY, Aranda-Díaz A, Jain S, Yu FB, Meng X, Wang M, Iakiviak M, Nagashima K, Zhao A, Murugkar P, Patil A, Atabakhsh K, Weakley A, Yan J, Brumbaugh AR, Higginbottom S, Dimas A, Shiver AL, Deutschbauer A, Neff N, Sonnenburg JL, Huang KC, Fischbach MA. 2022. Design, construction, and in vivo augmentation of a complex gut microbiome. Cell. 185(19):3617-3636.e19. doi: 10.1016/j.cell.2022.08.003
    • Arthofer P, Delafont V, Willemsen A, Panhölzl, F, and Horn M. 2022. Defensive symbiosis against giant viruses in amoebae. Proceedings of the National Academy of Sciences, 119(36). doi: 10.1073/pnas.2205856119.

1. Paper Abstracts

1.1. Snippet paper; discussion starts at 4:26 minutes

The Most Interesting Things (according to students)

  • When gnotobiotic (germ-free) mice are colonized with a set population of bacteria the community of bacteria within the mice stabilizes itself. The various strains within the community will remain at roughly the same proportions regardless of how many times the experiment is repeated.
  • When microbes from the human gut were added to the stabilized gut community, very few additional species were able to colonize. This included strains we think of as gut bacteria, like E. coli, which was not part of the stable community.

The abstract cannot be copied due to licensing restrictions. Please see licensing information and links to the article at the journal’s web page and/or PubMed in Section 8.1.

1.2. Main paper; discussion starts at 28:38 minutes

The Most Interesting Things (according to students)

Bacterial symbionts inhibit the replication of viruses when the two cohabitate in the amoeba Acanthamoeba.

“Protists are important regulators of microbial communities and key components in food webs with impact on nutrient cycling and ecosystem functioning. In turn, their activity is shaped by diverse intracellular parasites, including bacterial symbionts and viruses. Yet, bacteria–virus interactions within protists are poorly understood. Here, we studied the role of bacterial symbionts of free-living amoebae in the establishment of infections with nucleocytoplasmic large DNA viruses (Nucleocytoviricota). To investigate these interactions in a system that would also be relevant in nature, we first isolated and characterized a giant virus (Viennavirus, family Marseilleviridae) and a sympatric potential Acanthamoeba host infected with bacterial symbionts. Subsequently, coinfection experiments were carried out, using the fresh environmental isolates as well as additional amoeba laboratory strains. Employing fluorescence in situ hybridization and qPCR, we show that the bacterial symbiont, identified as Parachlamydia acanthamoebae, represses the replication of the sympatric Viennavirus in both recent environmental isolates as well as Acanthamoeba laboratory strains. In the presence of the symbiont, virions are still taken up, but viral factory maturation is inhibited, leading to survival of the amoeba host. The symbiont also suppressed the replication of the more complex Acanthamoeba polyphaga mimivirus and Tupanvirus deep ocean (Mimiviridae). Our work provides an example of an intracellular bacterial symbiont protecting a protist host against virus infections. The impact of virus–symbiont interactions on microbial population dynamics and eventually ecosystem processes requires further attention.” (Arthofer et al 2022, no changes)

2. Vision and Change Core Concepts and 2024 ASM Fundamental Statements

Snippet Main
Vision and Change Topics
  • Microbial Ecology (V&C_ME)
  • Impact of Microorganisms (V&C_IM)
  • Microbial Ecology (V&C_ME)
  • Impact of Microorganisms (V&C_IM)
ASM Fundamental Statements
  • Fundamental Statement 20 (ASM_20): Microbes are ubiquitous, found in diverse and dynamic ecosystems, where they use available resources and often form complex communities.
  • Fundamental Statement 22 (ASM_22): Most microbes interact with hosts in beneficial or neutral ways, with a minority having a detrimental impact on their host.
  • Fundamental Statement 19 (ASM_19): Non-cellular infectious agents, such as viruses, prions, viroids, and satellites, are dependent on host cell processes in order to replicate.
  • Fundamental Statement 22 (ASM_22): Most microbes interact with hosts in beneficial or neutral ways, with a minority having a detrimental impact on their host.
  • Fundamental Statement 29 (ASM_29: The extent of microbial damage can be minimized by host-derived and external factors, including the microbiome, antibiotics, and immunity.

3.  Potential Learning Objectives for the Podcast

The student will be able to: Paper1 Order2
  • Recall the use and importance of germ-free animal models in microbiome studies.
  • Identify the benefits of whole metagenomic sequencing approaches of analyzing microbial communities compared to 16S rDNA next generation sequencing approaches.
  • Identify how mice colonized by a synthetic microbiome compare to from mice colonized with a human fecal transplant.
 S L
  • Hypothesize a probiotic treatment based on these study conclusions.
 S H
  • Differentiate the taxonomy of viruses, bacteria, and amoeba.
  • Recall the ecological significance of the interactions between protists, viruses, and endosymbionts.
 M L
  • Evaluate the relationships of Parachlamydia and Viennavirus with Acanthamoeba to characterize them as commensalism, parasitism, mutualism, or predation.
 M H

1 Papers: Snippet (S) or Main (M)

2 Learning Objectives: Lower Order or Higher Order (H)

4. Techniques Described (with Time Stamps)

Here is a link to a bio-dictionary that has many, but not all definitions if you need a definition: Explore Biology Bio-Dictionary

4.1. Snippet Paper

  • Cell Culture (9:43–10:50): These are methods for growing cells in the laboratory.  Here, bacterial strains commonly found in the human gut-microbiome were grown.
  • Optical Density (10:06–11:00): This is a method for quantifying cells in a liquid culture using spectroscopy in the visible range.  The experimental community composed of varying bacterial strains was prepared and normalized using optical density.
  • Whole Genome Sequencing and Metagenomics (11:00–12:00): This is a method of using environmental DNA samples  to re-construct genomes of all species presnt in a mixed sample.  Here, whole genome sequencing, instead of 16S rDNA sequencing, was used for taxonomic identification due to the close relatedness of these bacteria.

4.2. Main Paper

  • Confocal Microscopy (36:58–38:10): This is a microscopy method that can generate high resolution 3-dimensional images.  It was used to obtain pseudo-colorized images of amoeba  with their symbiont Parachlamydia and show viral replication factories when infected with Viennavirus.
  • Quantitative Polymerase Chain Reaction (qPCR) (42:00–42:45): This is a molecular method to quantify DNA produced during the PCR reaction.  It is a method to quantify the DNA present in an original solution.  Here, qPCR was used to quantify the number of viral particles within the amoeba following infection.

5. Connections to General Microbiology Processes/Concepts

5.1. Snippet Paper

  • Gut Microbiome (4:30–6:00; 12:30–13:50, 15:27–22:10): Researchers used the most common strains of bacteria found in the human gut microbiome to build the microbiome for gnotobiotic (germ-free) mice.
  • Bacterial Cell Culture (8:50–10:57): The 104 bacterial strains identified from the human gut microbiome were grown in Chopped Meat Broth and in Mega Medium to form hCom1 community.
  • Metagenomics (10:59–11:40): Metagenomics was used to identify bacterial organisms down to the sub species level, enabling researchers to analyze and select bacterial strains for their study.

5.2. Main Paper

  • Symbiosis (32:20–35:00; 40:50–42:45): Mutualistic relationships are observed between Parachlamydia and Acanthamoeba.
  • Endocytosis (34:59–35:10): Isolated symbionts were introduced to the amoeba and were engulfed via endocytosis.
  • Viral Replication (35:35 -36:50): Viral replication in viral factors is observed inside of infected amoeba.

6. Podcast Questions

  1. What distinguishes “germ-free mice” from conventionally raised mice, making them a crucial tool in microbiome research? 
    1. They have a highly simplified and defined natural microbiome.
    2. They are genetically engineered to resist bacterial colonization.
    3. They are raised in a sterile environment and lack any microorganisms.
    4. They possess a robust immune system that eliminates all microbes.
  2. Which common approach(es) did the researchers use when conducting microbiome studies using germ-free mice? [pick all that apply]
    1. Colonizing them with defined microbial communities for controlled experiments.
    2. Observing how a spontaneous and complex microbiome developed in a sterile environment.
    3. Introducing single human pathogens to study their virulence without microbial competition.
    4. Using them to screen for novel antimicrobial compounds produced by commensal bacteria
    5. Colonizing them with complex samples from other hosts for controlled experiments.
  3. Researchers performed metagenomic whole genome analysis instead of 16S rDNA gene sequencing to identify and quantify the bacterial species. What was the main reason for this choice?
    1. Many organisms were so closely related that 16S rDNA sequence could not differentiate them.
    2. Metagenomics whole genome sequencing is now faster and cheaper than 16S rDNA sequencing.
    3. Sequencing 16S rDNA distinguishes fungi, so it cannot be used for bacterial identification.
    4. Metagenomic whole genome sequencing provides information about integrated viral prophages.
  4. How did the stability of the hCom1 synthetic microbiome compare to the augmented hCom2 community when repeatedly challenged with new human fecal samples?
    1. hCom1 was stable with the microbial community remaining unchanged after challenge, but hCom2 became unstable due to continuous challenges.
    2. Both hCom1 and hCom2 were found to be unstable. The community populations required constant reseeding in order for the mice to remain colonized.
    3. hCom1 was unstable with the bacterial communities unable to remain in constant proportions, but hCom2 showed greatly improved stability after augmentation.
    4. Both hCom1 and hCom2 demonstrated remarkable stability, although hCom1 gained new species and hCom2 resisted challenges from new samples.
  5. The podcasters note the ecological aspects of “healthy” or “perfect” microbiome community formation.  What implications might this have for entrepreneurs trying to formulate and sell probiotic treatments?
    1. There will one bacterial composition for each gut microbiome-related disease, so many disease-specific probiotics will be needed.
    2. There will be many bacterial compositions that will work as probiotics so long as bacteria that fill specific niches are included.
    3. There will only be a few bacterial compositions that will work and probiotic treatment should be taken regularly to ensure stability.
    4. There is a single bacterial species that could treat most gut microbiome disorders, so single species probiotics are needed.
  6. True/False:
    1. ______ Viruses belong to the prokaryotic domain.
    2. ______ Bacteria belong to the prokaryotic domain.
    3. ______ Amoeba belong to the eukaryotic domain.
  7. The podcast mentions “viral factories” within the amoeba, where viruses construct new virions. This concept highlights which fundamental aspect of viral reproduction, distinguishing it from bacterial or protist reproduction?
    1. Viruses have a complex internal organelle system for self-assembly inside hosts.
    2. Viruses are acellular and use host cellular machinery/resources for their replication.
    3. Viruses undergo a special and unique form of binary fission inside the host nucleus.
    4. Viruses reside in the host golgi where they are exported by the secretion system.
  8. According to the podcast, what is a key ecological advantage conferred by the Parachlamydia endosymbiont to its Acanthamoeba host in environments rich in giant viruses?
    1. It significantly increases the amoeba’s growth rate, allowing it to outcompete other protists.
    2. It makes the amoeba’s cell wall impenetrable to all viral infections, regardless of virus type.
    3. It provides resistance to lytic attack by giant viruses, enhancing the amoeba’s survival.
    4. It enables the amoeba to synthesize essential nutrients that are otherwise unavailable.
  9. Based on the podcast, the relationship between Parachlamydia acanthamoeba and its Acanthamoeba host can best be described as _______ because________.
    1. Parasitism; the Parachlamydia slows the growth rate of the amoeba, which harms the amoeba, while also benefiting from it.
    2. Commensalism; the Parachlamydia benefits from the host resources, but the amoeba host is neither helped nor harmed.
    3. Mutualism;  the Parachlamydia benefits from the intracellular lifestyle, and the amoeba gains resistance to giant viruses.
    4. Predation; when attached to the amoeba cell surface, the Parachlamydia forms a pore and consumes the amoeba contents.

7. Figure Reading Exercises

The following are two figure reading exercises, one from the snippet paper (Figure 7DE) and one from the main paper (Figure 1E–J).

7.1. First Figure Reading Exercise

7.1.1. Learning Objectives

Students will be able to:

  • Identify key features in schematics, strip plots, and Kaplan-Meier survival plots.
  • Identify key elements of experimental design including controls and variable types.
  • Identify the purpose of the phylum dropout experiment.
  • Analyze the data to make conclusions about each synthetic microbiome’s effect on pathogenic infection and host survival.
  • Analyze the data to make conclusions about each microbiome group’s contribution to resisting pathogenic infection and effect on host survival.
Experimental Background (Cheng et al, Figure 7DE)

It is becoming more and more clear that the gut microbiome plays a significant role in overall health, a robust immune system, and a variety of diseases, including autoimmune disease and age-related diseases.  An important feature of a healthy gut microbial composition (microbiota) is its ability to resist colonization and infection by pathogens, Cheng et al. (2022) wanted to investigate if they could produce a synthetic gut microbiome community to protect mice against specific pathogens, such as the gut pathogen, enterohemorrhagic E. coli (EHEC).  To do this, the researchers used information from the Human Microbiome Study and the metabolism literature to build a complex, defined microbial community (hCom1) of 104 species representing the typical human microbiome.  They next examined metabolism and metabolite literature to add species that filled un-utilized niches, bringing the total species to 119 for their “augmented” microbiome (hCom2).  To determine the effect of each four major bacterial group in the synthetic microbiome (Actinobacteria, Firmicutes, Proteobacteria, Verrucomicrobia) they used four variations of hCom2, each consisting of the same bacterial community, but with missing all species from each phyla (ΔActinobacteria, ΔFirmicutes, ΔProteobacteria, or ΔVerrucomicrobia).  To test the ability of these microbiomes to alter infection with enterohemorrhagic E. coli (EHEC), germ-free (gnotobiotic) mice were colonized with one of four variations of the hCom2 group or the hCom2 group (panel D). Following a four-week incubation, the variations of the hCom2 communities within the mice were challenged with EHEC. The amount of EHEC in the feces of the mice was quantified by plating feces on selective medium at 2 and 6 days post EHEC infection (panel E, left). The horizontal bar indicates either the mean or median (it is not clear which), while the asterisks denote statistical significance of at least p <0.05.  Survival of each group was also compared; hCom2 and ΔFirmicutes survival is displayed (panel E, right).

  • The figure cannot be copied due to licensing restrictions. Please see licensing information and links to the article at the journal’s web page and/or PubMed in Section 1.0.

7.1.2. Questions

  1. How many Actinobacteria species are present in hCom2? What color circle represents the data for the hCom2 bacterial community that is missing the Actinoobacteria?
    1. blue; 10
    2. green; 40
    3. red; 65
    4. purple; 3
    5. orange: 1
  2. What does the CFU/g feces indicate in this experiment (panel E)?
    1. A CFU represents the number of bacterial cells in a single colony. It was used to measure the number of total number of microbes in a gram of mouse feces.
    2. A CFU represents a single microorganism that replicates to form a visible colony. It was used to measure the number of EHEC microbes in a gram of mouse feces.
    3. A CFU represents how my bacterial cells you will have to plate in order to see visible growth. It was used to determine the dilution needed for accurate measurements.
    4. A CFU represents a group of microorganisms that divides and forms a visible colony. It was used to measure the total number of microbes in a gram of mouse feces.
  3. What notation in the data panel (panel E) tells you the pairwise comparison being made is statistically significant, that is truly different?
    1. horizontal bar
    2. circles
    3. asterisks
    4. nothing
  4. What was the purpose of challenging the mice with EHEC bacteria after colonizing them with different microbiome compositions related to hCom2?
    1. The researchers wanted to see if EHEC could fill the niche of the missing bacterial group and remain present in the community.
    2. The researchers wanted to determine which groups of bacteria were essential for protecting the host against EHEC infection.
    3. The researchers wanted to determine the effects of the gut microbiome if the immune system was activated from EHEC infection.
    4. The researchers wanted to investigate how EHEC bacteria could alter the genetic makeup of the synthetic microbiome.
  5. Which gut microbiome is the control in this experiment? What is its purpose?
    1. ΔActinobacteria; it is a model to test how EHEC behaves when one group is removed
    2. ΔFirmicutes; it is a setup to observe how EHEC spreads with altered gut composition
    3. ΔProteobacteria; it is a design to measure EHEC impact when one phylum is missing
    4. ΔVerrucomicrobia; it is a comparison for removing a single species compared to a phylum
    5. hCom2; it is a comparison to show the level of EHEC resistance when all are present
  6. Based on the data presented (panel E), which phyla were necessary for the gut community to resist EHEC colonization at day 2 post infection? What is your evidence?
    1. Actinobacteria or Verrucomicrobia were necessary because when they were removed from the gut community EHEC levels increased at day 2.
    2. Actinobacteria was the only phylum necessary because when it was removed the level of EHEC colonization remained unchanged.
    3. Proteobacteria and Firmicutes were necessary because when they were removed from the gut community EHEC levels increased at day 2.
    4. Proteobacteria was the only phylum necessary because when it was removed from the community the mice died from EHEC infection by day 2.
  7. What effect did removing Firmicutes from the gut microbiome community have on mouse survival after EHEC infection?
    1. The mice without Firmicutes all died by day 12, while none of the mice with the full hCom2 community were killed.
    2. The mice without Firmicutes started dying on day 12, while none of the mice with the full hCom2 community were killed.
    3. The mice without Firmicutes showed similar levels of infection and death after EHEC infection as compared to full hCom2.
    4. The mice without Firmicutes showed initial sensitivity with some early deaths, but survival plateaued after day 3.

7.2. Second Figure Reading Exercise

7.2.1. Learning Objectives

Students will be able to:

  • Identify key features in experimental design, microscopy images, and line graphs.
  • Evaluate the data to draw conclusions about the effects of virus, bacteria, and infection schedule on viral and bacterial replication.
  • Evaluate the data to draw conclusions about the relationship among the virus, symbiont, and protist.
Experimental Background (Arthofer et al, Figure 1E-J)

Protists are primarily unicellular eukaryotes that are found in nearly every ecosystem and have important roles in food webs, nutrient cycling, and shaping microbial communities. They are impacted by a wide variety of intracellular parasites, including both viruses and bacteria. In this study, Arthofer et al 2022 characterize and investigate a protist they isolated from a sample of activated sludge, Acanthamoeba hatchettii, along with a giant virus, Viennavirus.  Viennaviruses are known to cause lytic infections in other Acanthamoeba species. The environmental A. hatchettii isolate appeared to be resistant to lysis by the Viennavirus and also contained an endosymbiotic bacterium (Parachlamydia).  To determine the triad relationship among the virus, symbiont, and protist, as well as investigate the symbiont as a viral protector, the researchers infected an A. castellanii, a surrogate for A. hatchettii, in two ways.  Infection was either simultaneous, with both virus and symbiont (panels E and F), or in sequence, with symbiont first and virus 12 hours later (panels H and I).  Fluorescent microscopy was used to visually inspect the co-localization of amoeba, viral factories, and symbiont/bacteria at 12 and 24 hours after infection (panels E and H; F and I, respectively). Virus abundance up to 54 hours post infection was quantified using quantitative polymerase chain reaction (qPCR) for simultaneous infection (panel G) and sequenced infection (panel J).  Please note that in these panels (panels G and J), both experimental groups are compared to infection with virus alone.

6-part figure including 4 microscopy images and two line charts.

Figure 1. “Coisolation of Viennavirus, P. acanthamoebae, and A. hatchettii and evidence for symbiont-mediated virus inhibition. … A. castellanii simultaneously infected with Viennavirus and the Parachlamydia symbiont: FISH images (E) 12 hpi and (F) 24 hpi and quantification of viral particles with qPCR (G). A. castellanii infected with the Parachlamydia symbiont 12 h before the addition of Viennavirus: FISH images (H) 12 hpi and (I) 24 hpi and quantification of viral particles with qPCR (J). In all FISH images amoeba cells appear in magenta, nucleus (N) and viral factories (VF) in yellow, and bacteria in cyan. Statistical analysis for qPCR was carried out with two-tailed unpaired Student’s t test (*P < 0.05 and ***P < 0.001).” (Arthofer et al. 2022, images and text cropped to include panels E-J only).

7.2.2. Questions

  1. Match the following colors to the viruses, bacteria, or amoeba they denote in the immunofluorescent images above. [1 =  Purple;  2 = Yellow;  3 =  Cyan (blue)]
    1. ______ Virus
    2. ______ Bacteria
    3. ______ Amoeba
  2. How do the experimental conditions differ between panel E and H?
    1. The cell in panel E was imaged 12 hours after infection with the Viennavirus while the cell in panel H was imaged 24 hours after Viennavirus infection.
    2. The cell in panel E was infected with the Parachlamydia symbiont for 12 hours while the cell in Panel H was not infected with the symbiont at all.
    3. The cell in panel E was infected with virus and symbiont at the same time, while the cell in panel H was infected with virus 12 hours after the symbiont.
    4. The cell in panel E was imaged 12 hours after infection with the symbiont only, while the cells in panel H was imaged 12 hours after symbiont infection.
  3. Which treatment (infection schedule and time) shows the lowest viral load based on the microscopy images?
    1. The group infected with the Parachlamydia symbiont  and Viennavirus simultaneously and imaged 12 hours after infection.
    2. The group infected with the Parachlamydia symbiont and Viennavirus simultaneously and imaged 24 hours after infection.
    3. The group infected with the Parachlamydia symbiont before the Viennavirus and imaged 12 hours after viral infection.
    4. The group infected with the Parachlamydia symbiont before the Viennavirus and imaged 24 hours after viral infection.
  4. What is the earliest time that there is a statistically significant difference in viral load for the samples using the simultaneous infection compared to virus alone (panel G)?  What notation in the graph tells you this?
    1. 24 hours; whiskers
    2. 12 hours; circles
    3. 24 hours; asterisks
    4. none; nothing
  5. What is the earliest time that there is a statistically significant difference in viral load for the samples using the sequenced infection compared to virus alone (panel J)?
    1. 6 hours
    2. 12 hours
    3. 54 hours
    4. none; nothing
  6. Based on all of these data, what conclusion can you make about the triad relationship of this virus, symbiont, and protist?
    1. The symbiont inhibits viral replication.
    2. The virus inhibits symbiont replication.
    3. The protist inhibits viral replication.
    4. This protist is not infected by either.
  7. Which of the following is a logical reason for the difference observed between simultaneous and sequenced infections with virus and symbiont?
    1. When the symbiont is added first it has time to replicate intracellularly so there are more bacteria present to interfere with viral replication.
    2. When the endosymbiont and virus are mixed in the tube, they interact inside the tube which allows bacteria to stop viral replication quickly.
    3. When the symbiont is added before the virus it depletes nutrients the virus needs to replicate, stopping viral replication soon after infection.
    4. When the endosymbiont and the virus is introduced simultaneously the virus infects the bacteria instead of the amoeba, stopping replication.

8. Paper Information and Licensing

8.1. Snippet paper

  • Cheng AG, Ho PY, Aranda-Díaz A, Jain S, Yu FB, Meng X, Wang M, Iakiviak M, Nagashima K, Zhao A, Murugkar P, Patil A, Atabakhsh K, Weakley A, Yan J, Brumbaugh AR, Higginbottom S, Dimas A, Shiver AL, Deutschbauer A, Neff N, Sonnenburg JL, Huang KC, Fischbach MA. 2022. Design, construction, and in vivo augmentation of a complex gut microbiome. Cell. 185(19):3617-3636.e19. doi: 10.1016/j.cell.2022.08.003
  • This article is not licensed for Creative Commons use; see the article’s copyright information. Thus, the abstract and figures cannot be copied here. Please see the article at the journal’s web page

8.2. Main paper

  • Arthofer P, Delafont V, Willemsen A, Panhölzl, F, and Horn M. 2022. Defensive symbiosis against giant viruses in amoebae. Proceedings of the National Academy of Sciences, 119(36). doi: 10.1073/pnas.2205856119 
  • This article is licensed for Creative Commons use using CC BY 4.0, which allows re-use and adaptation with proper attribution and notation of any changes. 

License

Icon for the Creative Commons Attribution 4.0 International License

Podcast Annotation and Resources in Microbiology Copyright © 2025 by Rebecca Seipelt-Thiemann; Nancy Boury; Gwendowlyn S. Knapp; Amaya Garcia Costas; and Patrick Armstrong is licensed under a Creative Commons Attribution 4.0 International License, except where otherwise noted.

Share This Book