Legionella pneumophila: Lollipop Proteins and the Wonders of Adhesion

By: Katrina Holbrook '26

In 1976, 2,000 veterans gathered for a convention at the Bellevue-Stratford Hotel in Philadelphia. After the convention, Legion members began to fall ill. Eventually, 182 people got sick; of those, 147 required hospitalization, and 29 died. This incident was surrounded by media attention and conspiracy theories, with people wondering if foul play was involved. But the real “Philly Killer” responsible for the 1976 outbreak of Legionnaire's disease was a gram-negative bacterium identified as Legionella pneumophila.

The cover of TIME magazine from Aug. 16, 1976, source.

While L. pneumophila mainly infects amoeba and protozoa, it is an opportunistic bacterium that can infect cells in our lungs when contaminated water is aerosolized and inhaled. It is an intracellular bacterium, meaning that it can infect and replicate within host cells. One area of research is how these bacteria infect host cells, like our lungs.

Adhesion is a crucial part of host cell infection. When L. pneumophila adheres to our lung cells, it can become resistant to flushing. This means that when we cough or produce mucus to clear our lungs, this bacterium won’t be flushed out.

The paper “The Legionella collagen-like protein employs a distinct binding mechanism for the recognition of host glycosaminoglycans” focuses on a protein that L. pneumophila uses to adhere to our lungs. I liked this paper because it combines computational and experimental approaches. As a data science and biology major, I find articles that combine both of my majors super exciting. I also enjoyed learning about the lollipop-shaped protein at the center of this paper!

At its core, this is a structure-function paper that explores the Legionella collagen-like (Lcl) protein. In biology, you will often hear that for proteins, structure determines function, but what exactly does that mean? Let’s check out the structure and function of the Lcl protein. Let's start understanding the Lcl structure by taking a look at what it looks like.

Figure adapted from Rehman et al. (2024). Figure C shows a micrograph of Legionella collagen-like protein with a 5-nm scale bar. The green arrows point to the C-terminal domains, and the grey arrows point to the collagen-like repeat regions. Figure D shows a diagram of Lcl anchored to the bacteria.

This figure shows the protein as both a micrograph and a diagram. The micrograph image of the protein was taken using a transmission electron microscope. Isn’t it amazing that using microscopy, we can see and take a picture of something so tiny! I also like how we can see the structures shown in the diagram so well in the micrograph. In the diagram, we can see that the protein has three regions shown in the yellow N-terminal helix, the grey collagen-like repeat region, and the green C-terminal domain. On the micrograph, the green arrows point to the C-terminal domains, and the grey arrows point to the collagen-like repeat regions.

The N-terminal helix, shown in the diagram in yellow, is what keeps the protein attached to the surface of the Legionella bacteria. This region acts as an anchor to the bacteria.

The collagen-like repeat (CLR) region, shown in gray in the diagram, has the gray arrow pointing to it in the micrograph. The CLR contains repeats of amino acids. Repeats are when the same pattern of amino acids is repeated. In this region, there are repeats of a stretch of 15 amino acids. What I thought was fascinating was that the number of repeats varies between L. pneumophila strains. For example, Philadelphia-1, a L. pneumophila laboratory strain that was isolated from the 1976 outbreak, had 19 repeats in the CLR region. Isolates from hot springs had 13 repeats. However, some of the L. pneumophila strains isolated and cultured from clinical and environmental samples had only 8 repeats. In the paper, the scientist hypothesized that the number of repeats a strain has is the minimum number of repeats the Lcl protein needs to remain stable and not unfold. In hot springs, at higher temperatures, strains needed more repeats than strains from cooler environments needed. This variability between strains shows how bacteria are able to adapt to their environments!

The CDT domain, shown in green in the diagram, has the green arrows pointing towards it on the micrograph. An interesting finding about the CDT domain was that it was found to be highly conserved. So unlike the CLR region, this domain does not vary very much between the different L. pneumophila strains.

I really liked the “lollipop” description of the protein, because not only does the protein look like a lollipop, it also has other similarities. Expanding on this analogy, the lollipop is stuck to the outside of the bacteria, with the N-terminal helix acting as an anchor and the CLR region acting as a stick. A “lollipop stick” holds the candy and provides stability and structure; the CLR region does the same thing here. The structure of the CLR region, the “lollipop stick”, changes between different bacterial strains that experience different conditions. Bacteria in hot water need a sturdier stick than bacteria in colder water.

But we haven’t gotten to the good part of the lollipop yet. I’m talking about the candy part of a lollipop, of course! The CTD domain is sticky just like how candy is sticky. The CTD domain sticks to glycosaminoglycans.

Our lungs have glycosaminoglycans (GAGs) in both the lung cell membranes and the extracellular matrix. The CTD domain of the L. pneumophila collagen-like protein can stick to the GAGs because the CTD has a large surface that is positively charged. The flat, positively charged patches of the CTD surface allow the protein to recognize and stick to the negatively charged sulfate groups of the GAGs.

Other Legionella species have Lcl homologs, meaning they have a similar protein to L. pneumophila’s Lcl protein. Interestingly, in other Legionella species, their Lcl homologs still help them stick to host cells, but likely do not have the same "sticky" large positively charged surface that the CTD domain has. Since the Lcl-CTD homologs likely do not have a large positively charged surface, they may bind to different GAGs than the ones Lcl binds to.

What I thought was most interesting about the binding between L. pneumophila’s Lcl and GAGs was what the scientists discovered using molecular dynamics simulations. Molecular dynamics is a type of computer simulation that is used to analyze how molecules move over time, so it can be used to observe the interactions between Lcl-CTDs and GAGs as time passes.

Figure adapted from Rehman et al. (2024). The Legionella collagen-like C-terminal domain is shown in green. With C4S, a glycosaminoglycan, shown using stick representation. A, B, and C are units of the Lcl-CTD trimer. The three most frequent binding clusters are shown.

The Lcl-CTD is made up of three units: A, B, and C. In the simulations, the scientists discovered that C4S, a GAG found in the lungs, can bind in multiple modes: the more common major binding mode and the less common minor binding mode. The two major binding modes shown above are the same, just rotated, since the three units are symmetrical. What was riveting about the simulation is that the GAG does not just bind to one of these sites and stay there; it repeatedly moves between these binding sites. Essentially, the GAG would bind to a site, then release and move to a different site. The GAG did not bind at two distinct, separate sites but moved around!

Protein binding is sometimes described as a lock and key, but this is an oversimplification. Proteins and GAGs are dynamic and can change shape, also called conformation, meaning there is not a single lock and key. This is where the dynamic binding of Lcl-CTD and C4S comes into play. Lcl-CTD can detach and reattach in the different binding modes, allowing it to sample and recognize C4S in multiple conformations. This ability to recognize C4S in different conformations helps increase L. pneumophila's ability to adhere to our lungs by allowing it to bind to C4S in different conformations!

Lcl is key to adhesion and invasion in L. pneumophila. Discovering more about the proteins involved gives us insights into adhesion as well as new drug targets, with Lcl being a possible target for antibody therapy. This important research also discovered more about an intriguing binding mechanism.

When the 1976 outbreak of Legionnaires' disease began, the cause was a mystery, but since then, science has allowed us not only to discover the bacteria behind the mystery illness but also to research even the small details of these tiny organisms, like adhesion. Science has allowed us to take a picture of something as tiny as a protein and model complex binding interactions computationally. So much has been discovered from 1976 to 2026. This paper is a celebration of how far science has come in the last 50 years, and it makes me wonder what mysteries we’ll solve in the next 50 years.


About the Author:


Katrina Holbrook ‘26 is a biology and data science major from Minnesota. She thinks proteins are super interesting and has done a past project trying to predict binding interactions computationally. Outside of class, she enjoys reading and spending time outside.