Hox genes encode a family of conserved transcription factors that play important roles in specifying regions of the body plan along the anterior-posterior axis. A new paper in Development introduces new approaches and provides further insights into the transcriptional mechanisms controlling Hox gene expression during vertebrate development. To hear more about the story behind the paper, we caught up with the first author Zainab Afzal and her PhD supervisor Robb Krumlauf, Professor at the Stowers Institute for Medical Research.
Zainab Afzal (L) and Robb Krumlauf (R)
Robb, can you give us your scientific biography and the questions your lab is trying to answer?
RK: I received a degree in Chemical Engineering from Vanderbilt University and worked for several years as a biomedical engineer. In pursuing my engineering research interests I fell in love with the field of developmental biology and obtained a PhD from Ohio State University studying coordinate gene regulation in fungi. I did postdoctoral training at the Beatson Institute for Cancer Research (Glasgow, UK) making flow-sorted human chromosome-specific libraries to aid in mapping disease loci. I was interested in using mice to model disease and development and did a second postdoc with Shirley Tilghman at the Fox Chase Cancer Center (Philadelphia, PA, USA) where, in collaboration with Ralph Brinster, we pioneered the use of transgenic mice to study gene regulation in development. I set up my own lab in 1985 at the National Institute for Medical Research in London (now the Francis Crick Institute) and served as group leader and Head of the Division of Developmental Neurobiology. In 2000, I became the Founding Scientific Director of the newly established Stowers Institute for Medical Research (Kansas City, MO, USA) and in 2019 transitioned to the role of Investigator and Scientific Director Emeritus.
Our research is aimed at understanding the regulatory information and molecular mechanisms that control patterning of the basic body plan in vertebrates. Our group studies the hindbrain and Hox genes as models for understanding patterning mechanisms and regulatory networks that control how the brain and craniofacial features are regulated in development, disease and evolution.
Zainab, how did you come to work in Robb's lab? What drives your research today?
ZA: I joined the Interdisciplinary Graduate Program in Biomedical Sciences program at Kansas University Medical Center. I remember being very lost trying to find a lab for my PhD research, but the program allowed for rotations at Stowers Institute. I really liked the environment at Stowers and was excited by the projects and people in Robb's lab. I was happy when he agreed to become my thesis mentor. In thinking about projects, I have always been passionate about RNA, and my Master's research involved working with micro-RNAs. Robb agreed that I could pursue a project studying non-coding RNAs in the Hoxb cluster. However, like many things in science, my initial RNA project ran into complex problems, and this required me to rethink my goals. I was excited about the idea of monitoring the transcriptional dynamics of Hox genes in vivo and to study how an enhancer shared by multiple genes impacts their activity. I was fortunate, because this was a challenging project which had not been done before in vertebrate embryos. Through fun collaborations with technical experts at the Institute we managed to optimize single-molecule fluorescence in situ hybridization (smFISH) methods to monitor and analyze nascent transcription in sections of mouse embryos. It was so exciting to be able to systematically visualize, at the level of individual cells, how shared enhancers activate Hoxb genes in wild-type and mutant embryos.
What was known about the coordination of Hoxb transcription before your work?
RK: Dynamic gradients of retinoic acid (RA) were known to directly activate Hoxb genes in the developing hindbrain, in part through a series of retinoic acid response elements (RAREs) embedded within and flanking the cluster. There was evidence that these RAREs were part of three enhancers which coordinate global Hoxb responses to RA by regulating multiple genes. However, we did not know how these shared enhancers regulated or coordinated activity at the level of transcription. The tight clustering of enhancers and transcription units, in the relatively small region spanned by the Hoxb complex, raised questions regarding target specificity and selective preferences of the shared enhancers and whether coordinate regulation occurs through co-transcriptional coupling of Hoxb genes. We wanted to address this question.
Can you give us a summary of the paper in a paragraph?
RK: In this study, by imaging patterns of transcription in vivo at the cellular level we found that Hoxb genes do not appear to be coordinately regulated by simultaneous co-transcriptional coupling of all or specific subsets of genes in the cluster. We predominately detected nascent transcription of only a single Hoxb gene in each cell, with a much lower frequency of simultaneous transcription of one or two other transcriptional units. Analyses in embryos with mutations in three RA-dependent enhancers which globally regulate Hoxb expression revealed that each enhancer differentially impacts patterns of nascent transcription within the cluster. Selectivity and competitive interactions between these enhancers play an important role in coordinating and robustly maintaining the proper levels of nascent transcription of Hoxb genes. Our data imply that rapid and dynamic interactions between these enhancers and their target promoters potentiate transcription of individual genes in coordinating the response to RA.
Were you surprised that there is very low co-regulated transcription of specific subsets of Hoxb genes that share RARE enhancers?
RK: Yes, I found this result really surprising! I thought that we would observe multiple Hoxb genes and non-coding RNAs with distinct transcriptionally coupled bursting patterns in most cells. I expected to see evidence for active or ongoing nascent transcription of all of the genes at some level in most, if not all, of the cells in the neural tube. The fact that we primarily observed bursting of only one gene at a time, with occasional co-bursting of one or two other genes, and that the number of cells undergoing active transcription at any given moment was relatively small, drove home the idea that the potentiation of nascent transcription from shared enhancers must be incredibly dynamic. This is so intriguing and really motivates me to understand the underlying mechanisms.
ZA: Yes, I definitely found this surprising. We were looking at the neural tube in the tail region of mouse embryos, where we know that all of these Hoxb genes have overlapping domains of expression. Hence, one would expect that at the level of individual cells all of these genes would be actively transcribed and we would frequently see co-localized patterns of nascent transcripts. But instead, we saw very few instances of co-localized transcripts, showing how dynamic the on/off switch for active transcription of Hoxb genes is under inputs by the shared enhancers.
How do you propose the 3D architecture of the Hoxb locus may influence the regulation of transcriptional dynamics?
RK: A key observation from our analyses is that the shared enhancers embedded within the Hoxb complex do not appear to have a preference for activating the immediately adjacent genes, indicating that spatial or 3D relationships between the enhancers and their target promoters is more important than linear proximity along the chromosome. This is consistent with the idea that looping may generate transcriptional hubs that physically organize the three RARE enhancers in a manner that facilitates dynamic competition or synergy with each other in interacting with and activating their target Hoxb promoters. This could provide a modulated on/off switch for nascent transcription of individual and multiple genes.
When doing the research, did you have any particular result or eureka moment that has stuck with you?
ZA: The first time I saw ‘spots’ inside cells while optimizing the smFISH method in mouse embryo sections was a eureka moment. I almost didn't believe it, but it was so clear and robust that I was definitely excited! I remember running to Stowers microscopy experts collaborating with us with a smile on my face and asking them to check to make sure these were real spots. We verified the spectra and then began our journey to optimize the smFISH protocol. Another eureka moment was figuring out how to automate quantification of nascent spots which we did by deep learning. This opened the opportunity to systematically visualize and quantify spots in a whole section and compare multiple sections in wild-type and mutant embryos.
And what about the flipside: any moments of frustration or despair?
ZA: Initially comparing results across different embryos and genetic backgrounds was very frustrating. To solve this problem, I started collecting embryos at a specific somite stage of development for all wild-type and enhancer mutants. I also ensured that the histological sectioning scheme was consistent across samples; defining exact regions along the anterior-posterior axis at which the tail was sectioned and standardizing how tissues were placed on the slides. We also had to be consistent in imaging tissue sections and ensuring all settings stayed consistent throughout experiments. While initially frustrating, working through ways to account for variables led to robust and reproducible results, which was very rewarding.
“While initially frustrating, working through ways to account for variables led to robust and reproducible results, which was very rewarding”
Zainab, what's next for you after this paper?
ZA: I developed a love for fundamental biology and transcriptional regulation while doing research in Robb's lab at Stowers. To pursue these interests further, I have now joined Nipam Patel's lab at the Marine Biological Laboratory (MBL) to study transcriptional control in Paryhale hawaiensis. These crustacean shrimps are transparent, and their cells align into a beautiful grid as the embryo develops, which makes them an ideal model system to visualize regulation of body segmentation at the level of gene expression.
Robb, where will this story take your lab next?
RK: The surprising results from Zainab's experiments imply that the spatial relationships between genes and regulatory elements in the Hoxb cluster are dynamic and rapidly changing in the nucleus in ways that impact their potential to co-regulate genes. We really want to understand the dynamics and properties of these spatio-temporal distance relationships in tissues and cells where the Hoxb genes are active or inactive to gain mechanistic insights into how they modulate and potentiate nascent transcription. Towards this end, we have started a new series of experiments to measure the dynamics of distances between Hox genes and the shared enhancers in vivo at the level of individual cells using multiplexed DNA-FISH probes spanning the entire cluster and flanking regions. We hope the combination of DNA and RNA smFISH approaches will allow us to examine how activity and distance correlate with each other and provide mechanistic insights.
Finally, outside the lab, what do you like to do in your spare time?
RK: I really enjoy hiking in the mountains, canoeing, canal boating, exercising and being active outdoors. These activities are balanced by my love for good food and wine!
ZA: I love learning about different cultures – this manifests in me watching K-dramas and reading a wide variety of books. I also enjoy trying out local coffee shops and restaurants. Recently, with the move to MBL, I have come to enjoy walks on beautiful beaches in the area and bicycling around town.
Stowers Institute for Medical Research, Kansas City, MO 64110, USA and Anatomy and Cell Biology Department, Kansas University Medical Center, Kansas City, KS 66160, USA.
E-mail: [email protected]