Junjie Hu is fascinated by the interplay of structure and function. His muse is the endoplasmic reticulum (ER), a sort of Swiss army knife of the cell that holds two distinct shapes and carries out several functions.

Studying the ER is a long way from Hu’s first biology experiment. As a high school student in Shanghai, he participated in an extracurricular project to grow seedless watermelons. The project didn’t yield fruit, but the experience helped confirm his interest in life science, a hot area for undergraduates at the time. He went on to study biochemistry at Shanghai’s Fudan University, but it wasn’t until his graduate studies at New York University that he found his calling. Hu had already completed three rotations—short stints in different labs—without finding a place to do his thesis research when he tried his hand in a structural biology lab led by physicist Stevan Hubbard.

Upon entering the Hubbard lab, Hu came face to face with one of the major challenges facing structural biologists. It can be difficult to purify proteins—the large molecules that do most of the work inside the cell—and even more difficult to coax them to form crystals, which can be used to determine the protein’s structure. But beginner’s luck was with Hu: He produced protein crystals during the few months of his rotation. And with that success, he found his lab and his future research specialty.

Hu’s Ph.D. research “totally changed how I think about what’s going on in the cell,” he says. “It’s not like I didn’t know proteins are three-dimensional, but to really think of a protein that way is not easy … after I did my Ph.D. in structural biology, I learned to envision proteins and related cellular events in a three-dimensional way.”

Hu’s success continued as he teased out details of some of the interactions involving proteins that respond to insulin. But he found himself drawn to research on the ER, an organelle that, among other things, produces proteins, fats, and natural steroids. Hu was attracted by the ER’s unique structure: While most organelles are more or less round blobs, the ER is one part maze of folded sheets and one part tangled network of tubules. Hu wondered how one organelle could contain these two different structures and how that unique combination of structures aids its function. Since the ER produces insulin, studying it seemed like the right step after graduation.

Hu heard from David Ron, whose lab was across the hall, that HHMI investigator and ER researcher Tom Rapoport at Harvard Medical School was looking for structural biologists. On the day Hu went for his interview at Harvard, Rapoport had just been notified of his election to the National Academy of Sciences. “I think I was lucky that his election put him in a really good mood that day,” Hu says, laughing.

Rapoport’s lab had recently identified two types of proteins, reticulons and DP1/Yop1, which help shape the tubular portion of the ER. Hu joined the lab and set out to purify and crystallize proteins from these families. Though he never did get crystals to form, he found something perhaps more exciting: When the purified proteins were reintroduced into lipids (the main ingredient of the membranes) in vitro, ER-like tubules appeared. Hu had shown how important reticulons and DP1/Yop1 proteins are to building ER tubules.

Hu next turned his attention to learning how those tubules fuse with one another to form a distinctive network. He showed that the mammalian protein atlastin is needed to form this network and that it has analogs in yeast and plants. But knocking out the yeast protein has little effect on cells, whereas certain mutations in the human atlastin protein cause hereditary spastic paraplegia (HSP), a neurological disease that progressively weakens the legs, and mutating the plant protein causes short and wavy root hairs. Hu realized that atlastin and its relatives—and, therefore, the ER tubule network—are most important in cells that have long protrusions, such as plants’ root hair cells and the nerve cells that become abnormal in HSP.

With two important discoveries under his belt, Hu finished his postdoctoral fellowship and started his own lab at Nankai University in Tianjin, China. There he continues to study atlastin, in part by comparing it with a related protein that governs the membrane fusion for another organelle, the mitochondria. He plans to investigate the structure of sheet-like ER soon. “The ER has been identified for more than half a century, but no one really knew why these particular shapes were generated until recently,” he says. “It’s a fundamental question of cell biology, and also it’s a rather new area where a lot can be learned.”