Nerdy Jobs: Protein Crystallization, College

Nerdy Jobs S1E1, Exploring the Magic of Protein Crystallization

Dr. Raven spends a day growing protein crystals with Dr. Sarah Bowman in Buffalo, New York. A robot sets out drops too small to see, and some drops grow crystals within hours. Dr. Bowman explains that X-rays aimed at a protein crystal help scientists learn the protein’s shape and design medicines.

A NASA scientist uses a pipette to add protein solution to a crystal-growing tray at her lab bench

Meet the scientist

Dr. Sarah E. J. Bowman

Job titleStructural biologist. Associate Investigator and Director of the National High-Throughput Crystallization Center at the Hauptman-Woodward Medical Research Institute (HWI) in Buffalo, New York, as listed in her 2023 American Crystallographic Association bio.
A workdayThe center has run crystallization screens since early 2000, for over 2,000 user labs and over 18,000 samples. Each sample is tested against 1,536 cocktails in 400-nanoliter drops under oil. Imagers photograph each drop at day 1 and at weeks 1, 2, 3, 4 and 6, with brightfield, UV fluorescence and second harmonic generation imaging. Her research group develops methods to crystallize macromolecules and detect tiny crystals. The group also pairs crystallography with spectroscopy to study metal-binding proteins tied to bacterial infections and neurological diseases.
TrainingB.A. in English literature and women’s studies, Cornell College, 1996. B.S. in chemistry, Metropolitan State College of Denver, 2005. Ph.D. in chemistry, University of Rochester, 2010. Postdoctoral research at MIT (2010 to 2016) and Los Alamos National Laboratory (2016 to 2017).

Black-and-white photo of an early lab machine for making X-ray diffraction patterns of crystals — Photography means “drawing with light.”

Your word of the day is

kris-tuh-LOG-ruh-fee

crystallo-clear ice, rock crystalfrom Greek krýstallos

-graphydrawing, writing, recordingfrom Greek -graphía

The science of crystals and how they form.

A lab robot arm moves plates that hold hundreds of tiny wells

Hauptman-Woodward’s screening lab tests each protein sample against 1,536 crystallization cocktails. From 2000 to 2010, the lab tracked the samples it screened for structural genomics groups, large projects that set out to solve structures for many proteins.

Make a guess before you read the paper. What percent of those samples led to a structure deposited in the Protein Data Bank (PDB)? Give your reasoning in one sentence.

Hint

List the steps between a purified protein and a solved structure. Each step can fail.

Edward Snell, one of the paper’s authors, works at an X-ray diffractometer at NASA in 1998

Read the paper

Luft, Snell and DeTitta, “Lessons from high-throughput protein crystallization screening: 10 years of practical experience,” Expert Opinion on Drug Discovery, 2011. The paper is free to read.

Joseph Luft, Edward Snell and George DeTitta ran HWI’s high-throughput screening lab, the center Dr. Bowman directs today. Each sample gets 1,536 cocktails in drops of 200 nanoliters of protein plus 200 nanoliters of cocktail under oil, photographed at day 1 and at weeks 1, 2, 3, 4 and 6. Table 1 of the paper counts the samples the lab received each year. For the structural genomics samples, Table 1 also counts the structures later deposited in the PDB. The authors note that structures reach the PDB some time after a sample passes through the lab.

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

NASA astronaut Michael Hopkins fills protein crystal plates with a pipette on the International Space Station

Table 1’s printed Total row counts 3,876 structural genomics samples from 2000 to 2010 and 485 structures deposited in the PDB. The yearly rows add up to more than these printed totals.

Using the printed Total row of Table 1, ____ of all structural genomics samples from 2000 to 2010 led to a PDB structure.

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

NASA astronaut Megan McArthur loads protein crystal plates on the International Space Station

In 2006, Table 1 counts 418 structural genomics samples and 68 structures.

In Table 1, ____ of the 2006 structural genomics samples led to a PDB structure.

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

Two sealed bags holding orange protein crystal growth plates on the International Space Station

Which year in Table 1 had the highest percent of structural genomics samples become PDB structures?

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

Edward Snell lines up a protein crystal for an X-ray study at NASA in 2000

The authors note that structures reach the PDB some time after a sample passes through the lab, and they wrote the paper in 2011.

Why might the 2010 percent in Table 1 run lower than the percent those 2010 samples would reach in later years?

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

Microscope photo of small, clear rectangular protein crystals on a dark background

The paper reports that 269 of 823 structural genomics targets, or 32.7%, grew crystals somewhere in the screen. For those 269 targets, an average of 3.58% of the 1,536 conditions showed a crystal hit, about 55 conditions per sample.

If a protein hits in only about 55 of 1,536 crystallization conditions, why does HWI’s screening lab test all 1,536?

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

In a 1,536-condition crystallization screen, what do the clear and precipitated drops tell scientists?

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

A computer model of a small plant virus built from protein pieces, solved from crystals grown in space

In Table 1, about one structural genomics sample in eight became a structure. In the episode, Dr. Bowman explains how structures of hemoglobin and of the main protease of SARS-CoV-2 help scientists.

Given how often the path from sample to structure fails, why do scientists keep using X-ray crystallography to learn how proteins work?

See the table

Year	All samples	Structural genomics samples	Structures deposited in the PDB
2000	227	13	0
2001	426	28	1
2002	749	105	12
2003	1,268	225	20
2004	1,833	703	59
2005	1,784	776	86
2006	1,886	418	68
2007	1,505	637	82
2008	1,185	568	86
2009	748	403	96
2010	910	504	75
Total	11,581	3,876	485

The X-ray station at a synchrotron where scientists study protein crystals, with a robot arm that loads samples

Table 1 tracks HWI’s structural genomics samples and structures from 2000 to 2010.

Which summary fits Table 1 of the Luft, Snell and DeTitta paper?

Two NASA scientists check a protein crystal on a diffractometer, a machine that aims X-rays at crystals

Which question is out of reach for Table 1 of the Luft, Snell and DeTitta paper?

Lesson complete

Today’s word

crystallography

kris-tuh-LOG-ruh-fee

The science of how crystals form and of their forms and internal structure.

Back to the lesson Start over

Up nextapiaryNerdy Jobs: Beekeeping

Sources

Raven Baxter, #NerdyJobs S1E1: Exploring the Magic of Protein Crystallization, YouTube, unlisted
Episode transcript, with times from the episode’s YouTube captions
Luft and others, Expert Opinion on Drug Discovery, 2011
Lynch and others, Acta Crystallographica Section D, 2023
Budziszewski and others, Journal of Visualized Experiments, 2023
Ghosh and others, European Journal of Medicinal Chemistry, 2024 (nirmatrelvir and the main protease)
RCSB PDB-101 Molecule of the Month: Hemoglobin (Dutta and Goodsell, 2003), Lysozyme (Goodsell, 2000), Coronavirus Proteases (Goodsell, 2020)
RCSB PDB-101, Methods for Determining Atomic Structures
American Crystallographic Association, Sarah Bowman candidate bio, 2023
Hauptman-Woodward Medical Research Institute, news release on Dr. Bowman, 2017
Vision and Change core concepts and competencies, as summarized in Brownell and others, CBE Life Sciences Education, 2014
Word origins from Dictionary.com: crystallography, crystal, -graphy, protein, diffraction, calligraphy, photography, cryo-, chromatography, chromato-, graph, graphic, phonograph, lysozyme, lysosome
Spanish and French spellings from Wiktionary

Nerdy Jobs: Protein Crystallization, College

Nerdy Jobs S1E1, Exploring the Magic of Protein Crystallization

Meet the scientist

Dr. Sarah E. J. Bowman

Your word of the day is

Make a guess before you read the paper. What percent of those samples led to a structure deposited in the Protein Data Bank (PDB)? Give your reasoning in one sentence.

Read the paper

Using the printed Total row of Table 1, ____ of all structural genomics samples from 2000 to 2010 led to a PDB structure.

In Table 1, ____ of the 2006 structural genomics samples led to a PDB structure.

Which year in Table 1 had the highest percent of structural genomics samples become PDB structures?

Why might the 2010 percent in Table 1 run lower than the percent those 2010 samples would reach in later years?

If a protein hits in only about 55 of 1,536 crystallization conditions, why does HWI’s screening lab test all 1,536?

In a 1,536-condition crystallization screen, what do the clear and precipitated drops tell scientists?

Given how often the path from sample to structure fails, why do scientists keep using X-ray crystallography to learn how proteins work?

Which summary fits Table 1 of the Luft, Snell and DeTitta paper?

Which question is out of reach for Table 1 of the Luft, Snell and DeTitta paper?

Lesson complete

Today’s word

Look at these again