Novel structure of cockroach allergen Bla g 1 has implications for allergenicity and exposure assessment

Abstract

Background

Sensitization to cockroach allergens is a major risk factor for asthma. The cockroach allergen Bla g 1 has multiple repeats of ~100 amino acids, but the fold of the protein and the biological function are unknown.

Objective

To determine the structure of Bla g 1, investigate the implications for allergic disease, and standardize cockroach exposure assays.

Methods

Natural Bla g 1 and recombinant constructs were compared by ELISA using specific murine IgG and human IgE. The structure of Bla g 1 was determined by X-ray crystallography. Mass spectrometry and NMR were utilized to examine ligand-binding properties of the allergen.

Results

The structure of a recombinant Bla g 1 construct with comparable IgE and IgG reactivity to the natural allergen was solved by X-ray crystallography. The Bla g 1 repeat forms a novel fold with 6 helices. Two repeats encapsulate a large and nearly spherical hydrophobic cavity, defining the basic structural unit. Lipids in the cavity varied depending on the allergen origin. Palmitic, oleic and stearic acids were associated with nBla g 1 from cockroach frass. One Unit of Bla g 1 was equivalent to 104 ng of allergen.

Conclusions

Bla g 1 has a novel fold with a capacity to bind various lipids, which suggests a digestive function associated with non-specific transport of lipid molecules in cockroaches. Defining the basic structural unit of Bla g 1 facilitates the standardization of assays in absolute units for the assessment of environmental allergen exposure.

INTRODUCTION

The first report on cockroach allergy dates back to 1964, when Bernton and Brown described an association between sensitization to cockroach allergens and asthma morbidity in New York.¹ Subsequent studies have confirmed this original finding, especially among lower socio-economic populations, but also in suburban middle-class homes of asthmatic children.^2–4 Rates of cockroach sensitization are highest in the Northeast U.S.A., reaching up to 81% in the Bronx, New York, with the highest allergen levels in high-rise apartments.³ A recent study showed that prenatal exposure to cockroach allergens was associated with a greater risk of developing sensitization.⁵ Allergic sensitization to cockroach is a risk factor for the development of asthma, as is sensitization to other indoor allergens from dust mite, mold, dog or cat.^6–8

A major breakthrough in understanding the role of cockroach allergy in asthma was the identification and measurement of cockroach allergens. Bla g 1 and Bla g 2 were the first two major cockroach allergens to be identified.⁹ Both are secreted in the cockroach digestive tract, and sensitization occurs by inhalation of fecal particles that carry the allergens and are released to the environment.¹⁰ Both cockroach allergens have been used as markers of allergen exposure. Exposure to low doses of Bla g 1, 2 U/g of dust or less, was determined to be a risk factor for sensitization, however the risk factor plateaus above 4 U/g of dust.^{2, 11} A seminal study by Rosenstreich et al. established 8U/g of dust as threshold levels of “morbidity due to asthma”, as part of a National Cooperative Inner-City Asthma Study.^{2, 11} However, the levels of Bla g 1 exposure in all these studies were expressed in arbitrary units, and should be converted to absolute protein values to make comparable assessments of exposure to different allergens.

Assays to measure cockroach allergen exposure were originally developed with monoclonal antibodies raised against cockroach extracts in the 1990s.¹² The cockroach extracts used for assay standardization were arbitrarily assigned a certain number of units (U) for each allergen (Bla g 1 and Bla g 2) in a fixed volume of extract.¹³ Subsequently, the molecular cloning of Bla g 2 revealed that this globular protein is stable with a well-defined structure, and this allowed the measurement of this allergen in absolute units based on amino acid analysis of purified Bla g 2 (1 Unit = ~ 80 ng).^{14, 15}

However, Bla g 1 is a more complex allergen.^{10, 13, 16, 17} Molecular cloning of Bla g 1 revealed an unusual primary structure containing ~100 amino acid repeats. The gene for Bla g 1 originated from a ~100 amino acid sequence that duplicated. The sequence of the second repeat diverged from the first, and this “duplex” subsequently multiplied to include many “duplexes” of the repeated motif. If we number the repeats consecutively, comparisons between consecutive repeats are 26–29% identical, whereas comparisons among odd numbered repeats or among even numbered repeats show 96–98% identity (Online Repository Figure E1).¹⁸ The number of repeats in clones reported for group 1 cockroach allergens are between four and fourteen. ^{10, 19, 20} In addition to variability in length due to genetic diversity, the allergen may appear 33–37, 28, 25, or 6 kDa depending on the method used for purification or visualization.^{13, 16, 21}. The different natural Bla g 1 molecular forms result from cleavage by trypsin-like enzymes present in the cockroach gut. ^{10, 13, 16} Among insects, Bla g 1-homologous proteins are found with a similar genetic structure (Online Repository Figure E1).^{18, 22}

Given the complexity of this allergen, the Bla g 1 units remained arbitrary until the allergen would be better characterized. The determination of the structure of Bla g 1, presented in this paper, reveals a basic structural unit comprised of two consecutive amino acid repeats. Knowledge of the structure of Bla g 1 allows the standardization of assays to measure this allergen in absolute instead of relative units. In addition, the structure provides novel information about the function of this protein in the cockroach gut, the nature of the repeats, and its allergenic potential.

METHODS

Constructs

Natural Bla g 1 exists in multiple molecular forms and cloning of the gene revealed repeated amino acid sequences.^{10, 16} In order to determine the three-dimensional structure of Bla g 1, the largest repeated unit, containing a “duplex” from Bla g 1.0101 consisting of two consecutive repeats, was selected for expression in E. coli and crystallization, based on previous observations that proteolysis of the protein generated a similar sized molecule.¹⁷ GFP was fused to the N-terminus of the allergen to stabilize the construct and facilitate crystallization; the construct was named rBla g 1-GFP (see details in Online repository).^{23, 24}

For standardization purposes, three Bla g 1 preparations were tested containing Bla g 1 molecules of different lengths. Two were expressed by methanol induction in Pichia pastoris as previously described, and were named rBla g 1-PP.²⁵ These two preparations contained different molecular forms, resulting from expression of the same Bla g 1.0101 clone containing two duplexes (accession number AF072219), under different conditions (presence or absence of antibiotic zeocin for Lots 34074 and 33045, respectively). In addition, an isolated rBla g 1 construct of repeats 1 and 2 (rBla g 1-EC) was obtained by cleavage with TEV protease from a fusion with glutathione-S-transferase (GST) expressed in E. coli, (see details in Online repository). The reference standard containing natural Bla g 1 was a standard prepared from a Blattella germanica frass extract and used for ELISA that contains 10 U/mL of Bla g 1,¹³ (INDOOR Biotechnologies, Inc., Lot # 32023). The protein content of the three Bla g 1 preparations was quantified by amino acid analysis. Briefly, samples were hydrolyzed (6N HCl, 110C, 24hr), and the resulting amino acids were separated on a strong cation exchange column, detected with a secondary reaction with ninhydrin and quantified against a known standard run in the same sequence.

For NMR and MS studies, the Bla g 1 constructs rBla g 1-EC and rBla g 1-PP Lot 34074 were used. Natural Bla g 1 (nBla g 1) was purified from cockroach frass (debris and feces produced by the cockroach) using the anti-Bla g 1 mAb 10A6 and a similar protocol to that described previously for purifying nBla g 2 from frass.^{10, 25}

Antibody binding to nBla g 1 and rBla g 1 constructs

IgE antibody binding to natural and recombinant Bla g 1-GFP used for crystallization was compared by ELISA using microtiter plates that were coated with anti-Bla g 1 mAb 10A6, as described previously.²⁵ Following incubation with the cockroach extract or the recombinant allergen, sera from cockroach allergic patients (n = 15) were added, and bound IgE was detected using biotinylated goat anti-human IgE. Details regarding the sera are in the Online Repository. Control sera were from a non-allergic patient and two mite allergic patients. An IgE standard curve was obtained with anti-Der p 2 mAb, natural Der p 2, and chimeric antibody 2B12-IgE, with a dynamic range of 0.5–250 ng IgE/ml.²⁶

For standardization purposes, a polyclonal rabbit anti-Bla g 1 was used for detection instead of IgE in an ELISA, as previously described.¹³ Dose-response curves (n =4 per Bla g 1 preparation) were performed to compare three rBla g 1 preparations with the standard containing natural Bla g 1. Curves were fitted using the MATLAB and the equivalence between relative and absolute units at the level of the EC50 was determined.

Structural and Biochemical Characterization

Details of the crystallography, mass spectrometric analysis, and NMR procedures are given in the online repository.

RESULTS

Recombinant Bla g 1 is comparable to natural allergen

The rBla g 1 construct used for crystallization was compared to the natural allergen for human IgE antibody binding. There was an excellent quantitative correlation between IgE antibody binding to nBla g 1 and rBla g 1-GFP using sera from cockroach allergic patients (n=15, r = 0.96, p<0.001) (Figure 1). These results indicated that the recombinant and natural allergens were equally recognized by IgE antibodies and that the GFP molecule used in the rBla g 1-GFP construct for crystallography did not interfere with IgE antibody binding.

Figure 1.

Correlation between IgE antibody binding to rBla g 1-GFP and natural Bla g 1, using sera from cockroach allergic patients measured by ELISA.

Bla g 1 has a novel three-dimensional structure

The structure of rBla g 1-GFP was determined by X-ray crystallography (PDB code: 4JRB, Figure E2 and Table E1 Online Repository). The structure of each 100 amino acid repeat of Bla g 1 is comprised of six short helices. The first five helices of each repeat form a planar pentagon, while the sixth helix is positioned above and across the pentagon face (Figure 2). Two consecutive pentagonal structures come together to form a duplex in which the interacting helices from each repeat run in opposite directions. The structures of the two repeats are remarkably similar (Cα RMSD 1.4 Å over the first five helices) given only 26% identity of the amino acid sequences. Attempts to identify a similar fold in the protein database were unsuccessful using the program DALI.²⁷ The top 40 protein structures from the DALI search had low significance scores (5–7) but each was manually examined for similarities and none matched Bla g 1, indicating this is likely a new fold family.

Figure 2.

Bla g 1 Structure. A) “Top View” shows the first 5 helices forming a pentagon. The dashed red arrow indicates residues with missing electron density between helix 6 and 1′. The hydrophobic cavity is indicated by red mesh. B) Topology diagram showing the interactions of the helices. Electron density in the cavity was modeled with PA (panel C) and short acyl chains, rendered as sticks in panel D. D) The “Side View” shows how the helices interact between the repeats.

Defining the allergenic unit

Based on the structure and the data in Figure 1, a duplex of two repeated sequences appears to define the minimum structural unit. The recombinant allergens expressed in E. coli (rBla g 1-EC) and P. pastoris (rBla g 1-PP, lots 34075 and 33045) were compared for murine/rabbit IgG antibody binding to the ELISA standard containing natural Bla g 1 in extract. Note that rBla g 1-PP has two of the structural units defined in Figure 2, rBla g 1-EC has one structural unit, and the nBla g 1 has been reported as differently sized molecular forms.^{10, 12, 16, 21} Parallel and overlapping dose-response curves were obtained for the three allergen preparations by ELISA, indicating equivalent antibody binding on a g/ml basis (Figure 3A). The equivalence between absolute units and relative units of the natural Bla g 1 in the standard extract was determined to be 1 Unit equals 104 ± 1 ng (n = 4 per preparation; 3 preparations). Figure 3B shows an SDS-PAGE analysis of the samples used for standardization. The antibodies equivalently recognized the different preparations despite seemingly different fragmentation patterns. Mass spectrometry analysis showed that the fragmentation patterns of rBlag 1-PP resulted from partial proteolysis between and within the duplexes in the unstructured loops between the repeats (Online repository Table 2, Figure E5).

Figure 3.

Standardization of nBla g 1 Units. A) Binding curves with rabbit polyclonal antibodies to rBla g 1-EC (green circles, n=4), rBla g 1-PP Lot 34074 (blue circles, n=4), rBla g 1-PP Lot 33045 (magenta circles, n=4), and nBla g 1from the standard cockroach extract (Lot # 32023, red circles, n=4). Fits to the curves to the determine the EC50 are shown for nBla g 1 (red line), and the recombinant constructs, which were fit as one experimental data set, n=12 (black line). B) Silver stained SDS-PAGE of protein samples used in the standardization. Note that the concentration loaded on the gel was ~1,000 fold greater than the highest data point in the ELISA. A small amount of uncleaved (GST)-Bla g 1 is apparent in lane 4 at 50 kDa.

Lipid Binding of Bla g 1

Another defining feature of the Bla g 1 molecule is that the duplex of repeats 1 and 2 enclose a large spherical cavity that is lined exclusively with hydrophobic residues. The cavity is approximately 3,750 ų (Figure 2). For comparison, in a survey of almost 600 enzymes the mean active site volume was 1072 Å^3.28 Within the cavity exists electron density consistent with long chain fatty acids. This fragmented density was modeled with six carbon chains of variable length as well as one generic phospholipid.

In order to confirm the ligand content in the cavity, Bla g 1 molecules from various sources were analyzed for lipid content by mass spectrometry and NMR. In the E. coli expressed Bla g 1 constructs (rBla g 1-EC), phosphatidylethanolamine (PE) and phosphatidylglycerol (PG) head groups with various lengths of saturated and unstaturated diacyl chains were found in abundance by MS (Online Repository, Figure E3). Based on the total ion count, the Bla g 1-EC sample contained 60% PG and 40% PE, while the E. coli lysate containted 36% PG and 64% PE. rBla g 1-EC had fewer of the cyclopropyl acyl chains than are commonly found in E. coli membranes. Figure 4A shows the ³¹P NMR spectra of the phospholipids derived from the insoluble fraction of E. coli lysate, from rBla g 1-EC, and from GST as a negative control (see Online Repository Methods). The ³¹P NMR spectrum of rBla g 1-EC showed a stronger peak intensity at the shift of PG versus PE, which is the opposite of the normal distribution in E. coli (Figure 4B). The MS and NMR data indicate that the Bla g 1 has a relatively greater affinity for the anionic PG relative to the other lipid species present in E. coli, although it is not so selective that it binds to only this type of lipid. As a negative control, the GST sample did not show any significant phospholipid peaks (Figure 4C). In summary, the data shows that Bla g 1 has an affinity for saturated or nearly saturated acyl chains and preferentially binds PG over PE from E. coli.

Figure 4.

Comparison of the phospholipid content using ³¹P NMR of A) E. coli extract (positive control), B) rBla g 1-EC, C) GST expressed in E. coli (negative control), D) P. pastoris extract, and E) rBla g 1-PP. The ³¹P NMR peaks were assigned based on literature values, relative abundance, and the MS data in this manuscript.

A similar MS and NMR analysis of rBla g 1 expressed in Pichia patoris (rBla g 1-PP) showed that rBla g 1-PP retains phosphatidylinositol (PI), phosphatidylserine (PS), and phosphatidylcholine (PC). An NMR comparison of the relative abundance of phospholipids in an extract of P. pastoris (Figure 4D) and the yeast-expressed allergen (Figure 4E) showed that rBla g 1-PP preferentially retains PC and PI and does not bind phosphatidic acid (PA) nor PE. To determine if phospholipid binding was relevant in the cockroach, natural Bla g 1 purified from cockroach frass was analyzed by mass spectrometry. Surprisingly, no phospholipids were detected in nBla g 1. Instead, the excreted allergen contains primarily the non-phosphorylated fatty acids: palmitate (C16:0), stearate (C18:0) and oleate (C18:1), data not shown.

Comparisons with Per a 1 and P. rapae NSP protein

Bla g 1 is the first structure in a family of proteins that has been identified in cockroaches (Per a 1, 70–72% identity), butterflies (P. rapae NSP, 20–30% identity), and other insects.²⁹ Based on the sequence alignments in the Online Repository Figure E1, residues on the molecular surface of Bla g 1 were categorized according to their similarity with Per a 1 and P. rapae NSP (Figure 5). This analysis predicts that the Per a 1 and Bla g 1 surfaces will be nearly identical (Figures 5A & B). Using the Bla g 1 structure, calculations of solvent accessible area showed that 87% of the surface of cockroach Group 1 allergens is likely to be identical or similar (Figure 5C). The most notable area of sequence differences forming a contiguous surface is located in the segment from helix 4 to helix 6 (Figure 5B). In contrast, a similar comparison between Bla g 1 and a distantly related protein from P. rapae showed that only 39% of the surface area was predicted to be identical or similar (Figure 5D). The sequence differences between Bla g 1 and the other proteins were not limited solely to the external surfaces, but were equally distributed among buried and surface exposed residues in Bla g 1.

Figure 5.

Surface residue comparison between Bla g 1, Per a 1 and P. rapae NSP. A) and B) show the solvent accessible surface of Bla g 1 in two different orientations colored by residue identity (white), similarity (cyan) and differences (hot pink) with Per a 1. Visible helices are labeled by number beneath the semi-transparent surface. C) A pie chart of the total solvent accessible surface categorized similarly. D) A similar comparison with the distantly related P. rapae NSP.

DISCUSSION

In this study we present an unusual example of how the determination of the molecular structure of an allergen, Bla g 1, allows progress in the assessment of allergen exposure associated with respiratory symptoms. Over the years, measurement of environmental exposure to cockroach allergens has been essential to risk assessment for inner city asthma studies.^{2–4, 11, 30} Allergen measurements have also been important for integrative pest management to reduce allergen exposure and improve asthma.^31–33 Regarding diagnosis, a high variability in Bla g 1 and Bla g 2 content has been found in commercial cockroach skin test extracts, with up to 200-fold difference in Bla g 1 levels (4.7 to 1085 U/ml; n = 6 commercial extracts).³⁴ However, in all the previous studies, the levels of allergen exposure were expressed in relative units.

A process of assay standardization in absolute units is presented for Bla g 1. This usually requires the comparison of the Bla g 1 standard used in the environmental assays with a known amount of purified allergen using a reliable protein quantification technique. Natural Bla g 1 is a very unusual allergen that exists in multiple molecular forms, resulting from trypsin cleavage of the protein (Online Repository).¹⁰ A molecule that breaks down into fragments would be typically difficult or impossible to standardize due to the consequent abrogation of antibody epitopes. The importance of the new structural motif for the quantitation of Bla g 1 is that it provides a rationale to understand why in solution the molecule preserves antibody epitopes and yet appears fragmented on a denaturing gel. The 3D structure of repeats 1 & 2 determined by X-ray crystallography is novel, and the basic structural unit has the shape of a capsule or bead with a large internal cavity that accommodates lipid ligands. This suggests that natural Bla g 1 might resemble beads on a string, attached by flexible linkers (Figure E5). The multiple molecular forms of Bla g 1 observed by SDS-PAGE resemble fragments of a broken necklace, each with different numbers of beads or half beads since the long flexible linker between the repeats is also vulnerable to proteolysis (Figure E5). The crystallographic and antibody binding results suggest that the capsule structure is likely maintained in solution even with limited proteolysis. This scenario is very different from a globular molecule that degrades into different fragments with a consequent loss of epitopes, which makes standardization impossible.

We showed by ELISA dose-response curves, that the behavior of different Bla g 1 preparations was very similar regarding antibody binding and comparable to the natural allergen despite a different molecular fragment composition. The basic structural unit showed a good correlation with the natural allergen regarding IgE antibody binding as also reported for the Pichia expressed allergen.¹⁷ In addition, the four bands of the Pichia expressed Bla g 1 were shown to bind IgE antibodies,¹⁷ which is consistent with the fact that allergic sensitization occurs by inhalation of fragmented natural Bla g 1.¹⁰ In summary, a similar fragmentation pattern occurs in natural and recombinant allergens leading to molecules that are relevant for patient IgE and monoclonal antibody binding.

One arbitrary unit (U) of Bla g 1 was calculated to be equivalent to 104 ng, as absolute units. Bla g 2, with a similar equivalence of 80 ng per unit, has been reported to be very potent, inducing IgE sensitization at exposure levels that are 10–100 fold lower than other common indoor allergens, such as dust mite and cat.³⁵ The threshold levels for risk of sensitization and asthma morbidity (2 and 8 Units/g of dust, respectively) would be equivalent to 208 and 832 ng/g of dust for Bla g 1. These values were comparable to levels of 160 and 640 ng/g of dust for Bla g 2, but lower than values reported for dust mite and cat, confirming the potency of cockroach allergens. In summary, these results indicate that Bla g 1 and Bla g 2 have a similar allergenic potential, which is high compared to other indoor allergens, and allow direct comparison of allergen exposure measurements for different allergens.⁶

The structure of Bla g 1 is the first to be described for this group of insect proteins. Electron density in the cavity of Bla g 1 resembled phospholipids, and this in turn led us to investigate the internal ligands in Bla g 1 by MS and NMR. Both methods indicated phospholipids with various head groups could be retained by Bla g 1. MS analysis primarily identified saturated or monounsaturated acyl chains, as well as a few cyclopropyl acyl chains. Bla g 1.01 appears partially selective for certain phospholipid head groups including PC, PI, and PG. Promiscuous binding of lipids would be consistent with the idea of Bla g 1 being important for general nutrient uptake in cockroach: 1) the allergen is expressed in the midgut and excreted in fecal particles;^{10, 36} 2) the expression of Bla g 1 increases with food intake;³⁷ 3) knockdown of Bla g 1 expression using RNAi blocked nutrient absorption;³⁸ 4) homologous proteins in mosquitoes (ANG12 and AEG12) are up-regulated in the midgut after a blood meal,³⁹ and 5) AEG12 was localized by immuno-gold staining to microvilli in the gut.⁴⁰ Based on our results it is conceivable that Bla g 1 may be involved in the uptake of dietary lipids. The binding of certain lipids (palmitate, stearate and oleate) and not others to Bla g 1 from cockroach frass can be interpreted as a selective removal and excretion of these lipids into the feces, to facilitate retention and absorption of other, more nutritive compounds. Alternatively, homologous butterfly proteins allow the larvae to excrete toxic nitriles as an adaptation against the major chemical defense of their host plants.^{22, 41} In either case, the structure and data shown here suggests this protein fold could accommodate a variety of hydrophobic ligands for various physiological purposes.

A common feature of many allergens is the ability to bind hydrophobic ligands, which are potent stimulators of the innate immune system.⁴² The lipids that were found by MS in the cockroach frass are fatty acids that are known to activate TLR-4 and TLR-2.⁴³ Low-level stimulation of the TLR-4 pathway by LPS was shown to drive the immune response towards Th2,^{44, 45} and TLR-2 stimulation in the presence of allergen enhances Th2 responses.^{46, 47} Lipid adjuvants, like those found here associated with Bla g 1, can strongly influence the innate immune response and subsequently skew the adaptive response towards allergy. However, this remains to be demonstrated for nBla g 1 and fatty acids.

There is currently very little literature on specific epitopes of group 1 cockroach allergens except for two small scale studies with peptides of Per a 1.^{48, 49} Based on the ~70% sequence identity between Per a 1 and Bla g 1, the homologous allergen from the American cockroach is expected to have a similar fold and similar ligand specificity, and is unlikely to be a membrane spanning protein as was previously suggested.⁵⁰ Looking at the unique fold of Bla g 1, there are very few long stretches of amino acids that are significantly buried because there isn’t a canonical protein hydrophobic core. The hollow spherical shape exposes many more residues as potential IgE epitopes compared with a globular protein of similar mass, hence almost the entire sequence is potentially antigenic. The structure presented here should be very useful for future analysis of the antigenic surface and the design of Bla g 1 molecules for safer and more effective immunotherapeutics. It also provides insight into the digestive role of Bla g 1, with the potential adjuvant effects of lipids in allergenicity. The Bla g 1 structure facilitated measurement of the conversion factor from the previous arbitrary units to absolute units for the standardization of assays to assess allergen environmental exposure and allergen content in extracts used for diagnosis. These measurements are essential for understanding allergen levels associated with risk of sensitization and asthma morbidity, and for reducing allergen exposure and managing disease.

Table I.

Crystallographic data statistics
data set	rBla g 1-NGFP (high)	rBla g 1-NGFP (low)
unit cell	a=91.77 Å, b=141.92 Å, c=93.47 Å; α=β =γ= 90°	a=92.67 Å, b=141.62 Å, c=93.53 Å; α=γ= 90°, β =90.4°
Space Group	C2221	C2
Resolution (Å)	50.0 – 2.4	50–2.8
# of observations	165,823	103,638
unique reflections	23,505	29,002
Rsym(%)(last shell)1	8.9 (45.8)	6.8 (25.1)
I/σI (last shell)	14.9 (2.5)	11.6 (2.9)
Mosaicity range	0.69–0.93	0.27–0.58
completeness(%)(last shell)	99.3 (94.0)	97.8 (79.4)
Refinement statistics
Rcryst(%)2	19.8
Rfree(%)3	24.6
# of waters	100
Overall Mean B (Å2)	58.2
Average for: GFP	51.4
Bla g 1	67.6
ligands	75.3
water	51.8
r.m.s. deviation from ideal values
bond length (Å)	0.007
bond angle (°)	1.1
dihedral angle (°)	16.8
Ramachandran Statistics4
Residues in:
favored (98%) regions (%)	99.2
allowed (>99.8%) regions (%)	100

¹Rsym = Σ (| I_i − < I>|)/ Σ(I_i) where Ii is the intensity of the ith observation and <I> is the mean intensity of the reflection.

²Rcryst = Σ|| Fo| − | Fc ||/ Σ| Fo| calculated from working data set.

³Rfree was calculated from 5% of data randomly chosen not to be included in refinement.

⁴Ramachandran results were determined by MolProbity.

Table 2.

Mass Spectrometry Analysis of rBla g 1-PP

Mass (Da)	Repeat(s)	Residues (Bla g 1.0101)
9811.4	1 or 4	23–112, or 318–403
9967.8	1	38–128
10099.0	4	322–411
10237.1	2	108–198
10284.2	2	133–225
10374.0	1,3, or 4	38–131, 221–315, or 321–412
10505.0	2 or 3	129–221, or 218–313
10666.3	3	224–320
10965.0	1 or 4	12–111, or 316–412
11260.8	1	19–121

Clinical Implications.

• Bla g 1 like many allergens binds hydrophobic ligands. It is thought that these ligands may contribute to allergenicity as adjuvants.

• The structurally defined basic allergenic unit provides absolute standardization for cockroach exposure assessment in asthma.

Abbreviations

MS

Mass Spectrometry

GFP

Green fluorescent protein

GST

Glutathione-S-transferase

TEV

Tobacco Etch Virus

rBla g 1-GFP

recominant Bla g 1 with an N-terminal GFP

rBla g 1-EC

recombinant Bla g 1 expressed in E. coli

rBla g 1-PP

recombinant Bla g 1 expressed in Pichia pastoris

PA

Phosphatidic acid

PG

Phophatidylglycerol

PE

Phophatidylethanolamine

PC

Phophatidylcholine

PS

Phophatidylserine

PI

Phophatidylinositol

LPC

Lsyophophatidylcholine

CL

Cardiolipin

NSP

nitrile specifier protein

PBS

phosphate buffered saline

TLR

Toll-like receptor