Plasmodium falciparum serology: A comparison of two protein production methods for analysis of antibody responses by protein microarray

Abstract

The evaluation of protein antigens as putative serologic biomarkers of infection has increasingly shifted to high-throughput, multiplex approaches such as the protein microarray. In vitro transcription/translation (IVTT) systems–a similarly high-throughput protein expression method–are already widely utilised in the production of protein microarrays, though purified recombinant proteins derived from more traditional whole cell based expression systems also play an important role in biomarker characterisation. Here we have performed a side-by-side comparison of antigen-matched protein targets from an IVTT and purified recombinant system, on the same protein microarray. The magnitude and range of antibody responses to purified recombinants was found to be greater than that of IVTT proteins, and responses between targets from different expression systems did not clearly correlate. However, responses between amino acid sequence-matched targets from each expression system were more closely correlated. Despite the lack of a clear correlation between antigen-matched targets produced in each expression system, our data indicate that protein microarrays produced using either method can be used confidently, in a context dependent manner, though care should be taken when comparing data derived from contrasting approaches.

Introduction

To date, the majority of malaria serologic studies have focussed on antibody responses to a small number of well-characterised, highly immunogenic Plasmodium falciparum antigens that have proven to be reliable markers of exposure to infection [1–8]. However, P. falciparum expresses more than 5000 proteins, each a potential antibody target [9, 10]. Advances in technology have led to the development of new assay platforms that allow proteome scale investigation of antibody responses, such as the protein microarray [11, 12]—boasting significantly greater experimental throughput than more classical monoplex methods (e.g. ELISA) [13, 14]. The ability to simultaneously interrogate large numbers of putative targets, using low volumes of sample, significantly increases the rate at which an individual’s antibody responses to antigens can be characterised. As such, protein microarray based approaches to biomarker identification and humoral response profiling in malaria, and other infectious diseases, have been increasingly adopted [15–24].

One widely utilised form of the protein microarray is based on an in vitro transcription/ translation (IVTT) system [25]–where protein products are produced through a PCR, in vivo recombination cloning and an in vitro expression pipeline, before being printed onto arrays [15]. In principle, whole organism proteome microarrays can be fabricated simply and quickly, enabling analysis of all potential protein driven immune responses to a pathogen. Cell-free synthesis (CFS) is a technique first established over 50 years ago as a means to dissect the molecular mechanisms around protein expression. More recently, the technique has been used as a high throughput expression platform to explore a number of diverse biological processes [26, 27]. At its simplest, the approach utilises the crude extract containing the transcription and translation machinery from the cell, performing the process of protein expression without the constraints of the cell. This allows a wide variety of proteins to be expressed including those that would be deemed toxic if expression was attempted within the confines of the cell membrane [28]. CFS systems based on Escherichia coli (E.coli) are among the most widely used of the IVTT systems [27] and have helped to transform the narrative around a number of areas including biomarker discovery for infectious diseases [15, 29, 30]. Despite the widespread uptake of the approach there remain some issues around the technique. This includes significant heterogeneity of expression, leading some research groups to describe the mechanisms of the process as a “black box”. Therefore, the inherent heterogeneity between products is not assessed for every target making it difficult to normalise for reactivity between protein spots, which represent an impure mix of E. coli and target protein. In addition to the E. coli cell-free expression platform, other approaches have been employed in the characterisation of protein targets for immunological assessment. The wheat germ cell-free expression system in particular has also proven to be an important platform in the advancement of biomarker discovery and malaria vaccine research [31–34]. This is not the focus of the current study.

In contrast to the IVTT array methodology, the printing of purified proteins is cheaper and typically more quantifiable. Uniform amounts of product can therefore be incorporated into arrays, increasing confidence when comparing quantitative antibody responses between antigenic targets [35] and assessing relative immunogenicity. The process can be modified to support the scale up of recombinant proteins, and furthermore, affinity purification of protein targets reduces the risk of undesired background reactivity due to expression system components, and in part truncated proteins. However, the time required to produce panels of purified proteins is far in excess of the IVTT system, particularly for large numbers of targets, unless supported by an automated production platform [36–38]. For both the IVTT and purified protein E. coli systems, although the production of complex conformational proteins is possible it can sometimes be a challenge [39, 40]. These challenges are in part due to the expression of proteins foreign to the bacteria, the speed at which bacteria express proteins, only partially mitigated with a reduction in expression temperature; and the lack of essential molecular chaperones to aid correct folding/refolding of proteins [41–43].

Here we present a comparison between IVTT based and purified proteins on a single microarray. For clarity proteins produced using the IVTT system will simply be referred to as IVTT proteins, and those produced by conventional E.coli expression will be referred to as purified proteins. Matched malarial protein targets from each methodology were assessed for comparative reactivity in serum from Ugandan participant samples (n = 899) [44] to determine the suitability of each approach in the context of high-throughput profiling of serological responses to protein antigens.

Material and methods

Ethics statement

All serum samples were collected after written informed consent from the participant or their parent/guardian. The protocol for sample collection was reviewed and approved by the Makerere University School of Medicine Research and Ethics Committee (#2011–149 and #2011–167), the London School of Hygiene and Tropical Medicine Ethics Committee (#5943 and #5944), the Durham University School of Biological and Biomedical Sciences Ethics Committee, the University of California, San Francisco, Committee on Human Research (#11–05539 and #11–05995) and the Uganda National Council for Science and Technology (#HS-978 and #HS-1019).

Samples

Sera were originally collected as part of a comprehensive longitudinal surveillance study conducted in three sub-counties in Uganda (Walukuba, Jinja District; Kihihi, Kanungu District, and Nagongera, Tororo). The study design and methods have been previously reported and are described in detail elsewhere [44]. A sub-selection of samples (n = 899) was made from individuals across a breadth of recorded clinical episodes of malaria to ensure a range of sero-reactivity.

Protein targets

Purified protein expression.

Recombinant proteins were generated and expressed in Escherichia coli as glutathione S-transferase (GST)-tagged fusion proteins using previously described methods: PfMSP1-19 [45]; MSP1 block 2 [46]; ACS5, ETRAMP4 & HSP40 [19]; ETRAMP5 [19, 47]; EBA181 [48]; MSP4 [49]; MSP5 [50]; MSP7 [51]; and GAMA [52]. The exception to this was PfAMA1, which was expressed as a histidine tagged protein in Pichia pastoris [53]. Purification of the expressed proteins was performed using affinity chromatography (Glutathione Sepharose 4B (GE Healthcare Life Sciences) or HisPur Ni-NTA (Invitrogen) resins for GST and His tagged proteins, respectively). Protein concentration was assessed using the Bradford protein assay, with quality, and purity assessed by resolution on a 4–20% gradient SDS-PAGE.

IVTT protein expression.

An IVTT system was used to express proteins of interest as previously described [15]. Briefly, Plasmodium falciparum DNA (3D7 isolate) coding sequences were PCR-amplified and cloned into T7 expression vectors via homologous recombination. Target sequences were expressed at 21°C for 16h in E. coli-based, cell-free transcription/translation reactions, and products were printed onto arrays as un-purified, whole reaction mixtures.

Overview of compared IVTT and purified protein antigens.

We assessed antibody responses to protein targets mapping to eleven antigens (i.e. distinct gene products), each represented on the array by at least one IVTT and one purified protein target. Full details are in Table 1 and S1 Table. The number of purified protein targets varied according to availability, while the number of IVTT targets was dependent on the exon composition of each the gene sequence; multiple exon sequences were expressed as multiple protein targets based on exon delineation. Similarly, single exon gene sequences were generally expressed as a single protein. As a result, of the 11 antigens investigated, 8 were represented by >1 IVTT or purified protein target; 5 had >1 IVTT protein target (EBA181, HSP40, MSP1, MSP4 and MSP5) and 5 had >1 purified protein target (ACS5, ETRAMP4, ETRAMP5, HSP40 and MSP1). Near identical IVTT proteins (1 terminal amino acid difference in length) were produced independently and printed in parallel for two antigens: MSP4 and MSP5 as expression controls. Sequence information used in the design and expression of the purified E.coli proteins were generally smaller than the equivalent proteins expressed in the IVTT cell-free systems. This was done to limit the sequence length to below 1kb as expression of proteins larger that 1kb in E.coli can contribute to poor or failed expression yields [42, 43]. Truncation of target sequences was based on in silico mapping of each protein sequence to focus on regions of predicted immunogenicity based on the in silico analysis. Empty GST vectors were expressed and the purified GST used in background correction for proteins with this tag. His-tag vector was not expressed as it has proven impossible to express and purify the 6xhistidine tag in isolation.

[Figure omitted. See PDF.]

Protein microarray

Prior to printing, Tween 20 was added to purified proteins to yield a final concentration of 0.001% Tween 20. Arrays were printed onto nitrocellulose-coated slides (AVID, Grace Bio-Labs, Inc., Bend, OR, USA) using an Omni Grid Accent microarray printer (Digilabs, Inc., Marlborough, MA, USA). Alongside proteins of interest, buffer (PBS) and no-DNA (empty T7 vector reactions) were included as controls to allow for background normalisation of purified and IVTT proteins respectively.

Sample probing.

For analysis of antibody reactivity on the protein microarray, serum samples were diluted 1:200 in a 3 mg mL^-1 E. coli lysate solution in protein arraying buffer (Maine Manufacturing, Sanford, ME, USA) and incubated at room temperature for 30 min. Arrays were rehydrated in blocking buffer for 30 min. Blocking buffer was removed, and arrays were probed with pre-incubated serum samples using sealed, fitted slide chambers to ensure no cross-contamination of sample between pads. Chips were incubated overnight at 4°C with agitation. Arrays were washed five times with TBS-0.05% Tween 20, followed by incubation with biotin-conjugated goat anti-human IgG (Jackson ImmunoResearch, West Grove, PA, USA) diluted 1:200 in blocking buffer at room temperature. Arrays were washed three times with TBS-0.05% Tween 20, followed by incubation with streptavidin-conjugated SureLight P-3 (Columbia Biosciences, Frederick, MD, USA) at room temperature protected from light. Arrays were washed three times with TBS-0.05% Tween 20, three times with TBS, and once with water. Arrays were air dried by centrifugation at 500 x g for 5 min and scanned on a GenePix 4300A High-Resolution Microarray Scanner (Molecular Devices, Sunnyvale, CA, USA). Target and background intensities were measured using an annotated grid file (.GAL).

Data normalisation.

Microarray spot foreground and local background fluorescence data were imported into R (Foundation for Statistical Computing, Vienna, Austria) for correction, normalisation and analysis. Local background intensities were subtracted from foreground using the backgroundCorrect function of the limma package [54]. The backgroundCorrect function was then further applied to GST-tagged purified proteins, whereby background-corrected GST fluorescence was subtracted from background-corrected target fluorescence to account for any GST-specific reactivity in samples. All data were then Log2 transformed and the mean signal intensity of buffer and no-DNA control spots were subtracted from purified and IVTT proteins respectively to give a relative measure of reactivity to targets over background (S1 Fig) [20].

Results

Table 1 summarises the purified and IVTT protein targets for each antigen, with further detail in S1 Table. In brief, we assessed IgG antibody responses to 35 antigenic targets, derived from 11 well-characterised P. falciparum protein antigens (distinct gene products). Each antigen was represented by at least one IVTT and one purified protein target.

Magnitude of responses between expression systems

The magnitude of response to all protein targets was compared by antigen to evaluate differences in seroreactivity between IVTT derived and purified protein targets. As expected, responses varied significantly between antigens and between the protein targets mapping to each antigen.

Mean responses to all targets were compared by expression system (Fig 1) revealing a greater range of response to purified proteins (IQR Log2MFI = 3.88–6.40) than IVTT proteins (IQR Log2MFI 0.46–1.68), and a greater magnitude of response to purified than IVTT targets (p = <0.001). Similarly, the range and median intensity of individual antibody responses was found to be greater for purified proteins than their IVTT counterparts (e.g. AMA1—IVTT_1, median [IQR] Log2MFI = 1.66 [0.80–2.53]; Pure_1, median [IQR] Log2MFI = 7.92 [6.16–8.52]) for all targets (p = <0.001) except MSP1 Pure_2, which more closely reflected the level of reactivity to the two MSP1 IVTT targets (Fig 2).

[Figure omitted. See PDF.]

The mean magnitude of response of each protein target stratified by expression system, presented with median and interquartile range of all mean responses.

[Figure omitted. See PDF.]

All sample responses (n = 899) to all protein targets grouped by antigen, presented with median and interquartile range.

Correlation of responses between antigen matched targets

Considering all at least partially sequence matched IVTT and purified protein targets (i.e. excluding pairwise comparisons where purified protein sequence were completely non-overlapping with IVTT sequence for the same antigen) there was no evidence for a general correlation in mean response between expression platforms (Spearman’s rho (r_s) = 0.279, p = 0.23). Antibody responses to all protein targets for each antigen were therefore compared individually (representative example in Fig 3 and all antigens in S2 Fig). This allowed for comparison between sequence matching IVTT and purified protein targets (e.g. HSP40 IVTT 2 vs. HSP40 Pure 2), non-matching IVTT and purified protein targets (e.g. HSP40 IVTT 2 vs. HSP40 Pure 1), and matching or non-matching targets produced in the same system (e.g. HSP40 IVTT 1 vs. HSP40 IVTT 2). Correlations were highly variable (r_s = 1.00 to -0.045) though all but one (GAMA; r_s = -0.045, p = 0.17) demonstrated a degree of positive, if not always statistically significant, association.

[Figure omitted. See PDF.]

A representative example correlogram of multiple antigen-matched targets (left). Spearman’s rank correlation reported (r_s) and increasing blue colour scale indicates relative strength of correlation based on calculated correlations for all proteins included in this analysis. Protein schematic (right) represents amino-acid aligned representation of IVTT (green) and purified (orange) proteins to the full-length native protein (grey). Proteins in the correlogram and schematic are correspondingly aligned. Corresponding axes are adjacently below or to the left of each protein.

Multiple IVTT targets were produced for EBA181, HSP40, MSP1, MSP4 and MSP5. For all other than MSP5, non-sequence matching IVTTs were produced; correlation co-efficient for these targets were between 0.37 and 0.73 (S2 Fig). For EBA181 and MSP1, IVTT targets overlap by 17 amino acids–equivalent to a small peptide in terminal regions unlikely to cover immunogenic epitopes. As such, these targets were considered non-overlapping. For MSP4 and MSP5, duplicate IVTT protein products were generated for each gene, with each duplicate protein identical to the other except for the omission of one [N- or C-] terminal amino acid. These respective targets resulted in near perfect correlation of antibody responses (MSP4 r_s = 1.00, p = <0.001; MSP5 r_s = 0.94, p = <0.001). Multiple purified protein targets were produced for ACS5, ETRAMP4, ETRAMP5, HSP40 and MSP1—none of which overlap. Correlation between these purified protein targets in each antigen varied between 0.31 and 0.59 (S2 Fig).

For the 8 antigens with >1 IVTT or purified protein target, the greatest level of correlation was found between an IVTT and purified target in 4/8 instances; between two IVTT targets (IVTT-IVTT) in 3/8 instances; and between two purified targets (purified-purified) in 1/8 instances (S2 Table). Comparing correlations between antigen-matched IVTT and purified proteins only, overlapping targets correlate more highly than non-overlapping targets. Sample sizes were too low to test the significance of this trend within antigens (Fig 4).

[Figure omitted. See PDF.]

Targets with overlapping amino acid sequences are indicated by closed circles, compared to non-overlapping sequences indicated by open circles.

Discussion

Protein microarrays are a practical approach to the serological screening of large numbers of putative malaria antigen biomarkers. The throughput and flexibility of the microarray platform presents an opportunity to interrogate malarial antibody responses at a scale far exceeding traditional mono- or multiplex approaches, agnostic of predicted immunological targets. Here we have evaluated matched antigenic targets produced using two E. coli-based expression techniques–in vitro transcription/translation (IVTT), and purified, whole-cell recombinants–in the context of a protein microarray. We found that the magnitude of antibody responses to purified protein targets was generally higher than for their IVTT counterparts, and that correlation between protein target pairs at the individual serum sample level was variable and related to degree of sequence homogeneity between targets. Our findings warn against direct comparisons of microarray data from proteins produced in different expression platforms without careful cross-validation of sequences and allelic types. However, our data do provide support for the use of both IVTT and purified protein microarray platforms in the context of early-stage antigen biomarker identification to feed into experimental pipelines where candidate proteins may be interrogated by methods providing higher resolution analysis.

In building this study, we predicted that the magnitude of responses to IVTT products–which tended to be longer, often representing single exon sequences and therefore potentially containing more epitopes–would be greater than purified targets truncated based on species-specificity or domain boundaries which potentially represented fewer epitopes. Contrary to this prediction, we found that purified proteins captured a greater range and magnitude of responses (Purified, IQR Log2MFI = 3.88–6.40; IVTT, IQR Log2MFI 0.46–1.68; p = <0.001). The greater level of reactivity to purified targets may relate to differences in the amount of protein deposited on the array, where consistent and defined amounts of purified protein are spotted in contrast to the unquantified, and likely variable IVTT products. These findings recommend a degree of caution in interpretation of array data from two different platforms, for example: MSP5 showed the second highest mean MFI for any purified protein, but showed among the lowest mean MFI of any IVTT protein.

In addition to differences in the magnitude of mean responses to targets stratified by expression system, we observed a greater range of individual sample responses, stratified by antigen, to purified proteins than in sequence matched IVTT-expressed targets (e.g. AMA1—IVTT_1, median [IQR] Log2MFI = 1.66 [0.80–2.53]; Pure_1 (Pichia pastoris produced), median [IQR] Log2MFI = 7.92 [6.16–8.52]; p = <0.001). The P. pastoris AMA1 was included as a control for the evaluation of the production of a conformational protein. AMA1 is a complex structure comprised of three domains defined by three disulphide bonds. Production of AMA1 in P. pastoris has been fully characterised in terms of correct folding of the purified protein [53, 55] and this observation is likely a reflection of antibody reactivity to correctly folded (P. pastoris) and incorrectly folded AMA1 (IVTT). We acknowledge that a lack of correct folding in other purified and IVTT products may impact on epitope recognition by antibodies raised to native protein during infection. However, human antibody responses are composed of a polyclonal response to each antigen, which will include both confirmation and linear epitopes. Whilst questions remain about the appropriateness of using unfolded protein fragments in serological screens, such reagents remain the most widely utilised and efficient approach in this context at present.

Considering all antigenic targets together, we found no evidence of correlation in mean reactivity to sequence matched targets between expression systems (r_s = 0.28, p = 0.23). In the context of this study, this was not unexpected taking into account the differences observed in magnitude of response between IVTT and purified proteins, and that the length of native protein sequence coverage between IVTT and purified targets was highly variable. More broadly, it is perhaps less reassuring that matched targets derived from different expression systems lack more obvious relationships in antibody response than have been demonstrated in other studies [17, 56], though Kobayashi et al. report relatively similar results for a smaller number of targets expressed in E. coli (purified proteins) and IVTT systems specifically [30]. It is likely that protein concentration disparities between the two approaches are one of the drivers of this heterogeneity. However, without attempting to quantify the exact amount of protein generated in the small volume of IVTT reactions we are unable to address this here. Although in this current study targets grouped by antigen displayed highly variable correlations of response, it is encouraging that sequence matched proteins did generally display stronger correlations of response than non-sequence matched targets. Further, this may indicate the importance of capturing specific epitopes within expression sequences when producing antigens by either expression method.

Despite the lack of a clearly defined relationship between antigen-matched targets from the evaluated expression systems, we remain confident that microarrays utilising IVTT or purified recombinant proteins are able to produce compelling and biologically relevant data. Indeed, our data show age-dependent trends in antibody responses (typical of highly endemic populations) [1–3] irrespective of expression system (S3 Fig), lending weight to the applicability of either methodology in serological assays [57–70].

The IVTT system lends itself to microarray applications, as vast numbers of proteins, or even entire proteomes, may be produced at scale relatively quickly. However, for application to serology there is concern that expressed proteins are not quantified before printing, and that expression levels of product may vary considerably; product yield in bacterial-based IVTT systems is generally considered to be lower (typically ~1 mg mL^-1 or less) though higher protein yields have been reported [71, 72]. This has been shown to be due to an inherent heterogeneity with IVTT components, although this weakness is an area of active research [26, 73]. Similarly, it is important to acknowledge that the un-purified nature of printed reaction mixtures may mask, or otherwise adversely affect, the detection of antibody reactivity in a sample; Davies et al. report IVTT reaction compositions of 99% E. coli lysate to 1% target protein [15], though this will vary considerably, at scale, in practice.

In contrast to IVTT-based microarrays, printing purified protein allows a highly quantifiable approach to be taken. Affinity purification and dialysis of expression products substantially reduces the risk of background reactivity to bacterial components, and the simple determination of target protein concentrations allows defined quantities of product to be spotted, providing much greater confidence when comparing reactivity between targets. However, these advantages come at a substantial cost; the need for in silico analysis to design vectors, transfection procedures, expression and purification drastically slows the rate at which putative targets can be produced and screened. Shorter, epitope specific sequences may in theory be transposed from IVTT systems with a view to generating more granular serological screens, though we accept that truncated protein targets will in some cases favour linear B cell epitopes, while missing conformational epitopes. However, for measuring exposure to infection there is less importance on the targeting of confirmation epitopes than would be required for protective epitopes [57].

The primary benefit of the microarray platform is the ability to screen orders of magnitude more targets simultaneously than more standard serological assays. Our analysis shows that both IVTT and purified proteins can be successfully used to capture malarial protein-antigen specific antibody responses on a protein microarray. Although correlations of response between expression systems are not as strong as may have been expected, a number of acknowledged technical differences in the methods of protein production may account for this finding. In addition to the E. coli in vivo and IVTT systems utilised here, high-throughput wheat germ cell free systems have been successfully used to conduct large scale serological screens of putative antigen biomarkers [74, 75], alongside chemically synthesised peptide arrays [57, 62]. High-throughput mammalian and baculovirus expression systems have also been pioneered for the production of recombinant proteins [36, 76]. Differences in expression efficiency and the homology to native epitopes achieved by the assortment of available approaches likely have considerable impact on the capture of antibody from sample. This variability should be accounted for both in terms of choosing an experimental approach and comparative analysis between different methods. We suggest that further investigation of differences in seroreactivity to sequence-matched proteins derived from contrasting expression systems is needed to shed light on the parity between such data that is already widely published. It should also be noted that it is unlikely that any single expression platform will satisfy the demands of all recombinant expression projects due to varying importance such as protein folding, proteins activity (e.g. enzymes) and glycosylation. In addition, E. coli expression has the advantage of low cost, flexibility and easy scale-up.

Considering the data presented here more broadly, observed trends lend support to the utilisation of both IVTT and purified arrays depending on the objectives and context of hypotheses to be investigated. The strengths and weaknesses of each expression system should dictate the chosen approach on a case-by-case basis. For example, very high-density proteome level screening to identify ‘shortlists’ of candidate markers based on binary categorisation of seropositivity may be best achieved using IVTT systems. In contrast, smaller numbers of ‘shortlisted’ targets expressed as purified proteins may allow for more nuanced characterisation of antibody responses on a more continuous scale. As already described, the key limitation in the production of purified recombinants in our current expression pipeline is throughput. The adaption of our methods to increase the capacity of protein production would improve our ability to more widely mine the biomarker information derived from the IVTT platform. As such, we are currently exploring a number of existing approaches to address this methodological bottleneck [38, 77].

In summary, the IVTT protein microarray approach has proven to be a powerful, high-throughput, biomarker discovery platform with applicability across a range of infectious diseases. When combined with a cheap, scalable and flexible protein expression platform such as the E. coli in vivo expression platform we have the ability to mine potential diagnostic and vaccine related targets.

Supporting information

S1 Fig. Data normalisation processes for IVTT and purified protein spots.

After local background correction using the backgroundCorrect function from the limma package, purified protein spots were additionally corrected for possible GST reactivity by subtracting GST reactivity using the same function. After Log2 transformation, IVTT and purified proteins were normalised to background control spots of empty T7 vector and PBS buffer control spots respectively.

https://doi.org/10.1371/journal.pone.0273106.s001

S2 Fig.

Correlogram of multiple antigen-matched targets (left). Spearman’s rank correlation reported (rs) and increasing blue colour scale indicates relative strength of correlation based on calculated correlations for all proteins included in this analysis. Protein schematic (right) represents amino-acid aligned representation of IVTT (green) and purified (orange) proteins to the full-length native protein (grey). Proteins in the correlogram and schematic are correspondingly aligned.

https://doi.org/10.1371/journal.pone.0273106.s002

S3 Fig. Magnitude and range of response to IVTT and purified proteins, stratified by age.

All sample responses (n = 899) to all protein targets grouped by antigen, presented with median and interquartile range.

https://doi.org/10.1371/journal.pone.0273106.s003

S1 Table. Detail of expressed protein targets.

A key to the simplified nomenclature used for specific proteins in text is provided.

https://doi.org/10.1371/journal.pone.0273106.s004

(XLSX)

S2 Table. Correlation coefficient results for all protein pairs.

Protein targets are grouped by antigen, and all possible combinations within each antigen group are shown.

https://doi.org/10.1371/journal.pone.0273106.s005

(XLSX)

Acknowledgments

We thank all the study participants who participated in the original PRISM study (Program for Resistance, Immunology, Surveillance, and Modelling of Malaria in Uganda; East Africa ICEMR) from which the test samples were taken. The authors give thanks to James Beeson for the provision of purified recombinant protein EBA140 RIII-V, Ross Coppel for the provision of MSP4 and MSP5, and Tony Holder for the provision of GAMA and MSP7.

Citation: Oulton T, Obiero J, Rodriguez I, Ssewanyana I, Dabbs RA, Bachman CM, et al. (2022) Plasmodium falciparum serology: A comparison of two protein production methods for analysis of antibody responses by protein microarray. PLoS ONE 17(8): e0273106. https://doi.org/10.1371/journal.pone.0273106

About the Authors:

Tate Oulton

Roles: Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing

Affiliation: Department of Infection Biology, London School of Hygiene and Tropical Medicine, London, United Kingdom

https://orcid.org/0000-0002-3027-7380

Joshua Obiero

Roles: Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – review & editing

Affiliation: Department of Physiology and Biophysics, University of California, Irvine, Irvine, CA, United States of America

https://orcid.org/0000-0001-9741-512X

Isabel Rodriguez

Roles: Conceptualization, Investigation, Writing – review & editing

Affiliation: Department of Medicine, University of California San Francisco, San Francisco, California, United States of America

Isaac Ssewanyana

Roles: Resources, Writing – review & editing

Affiliations Infectious Diseases Research Collaboration, Kampala, Uganda, Makerere University College of Health Sciences, Kampala, Uganda

Rebecca A. Dabbs

Roles: Methodology, Resources, Writing – review & editing

Affiliation: Department of Infection Biology, London School of Hygiene and Tropical Medicine, London, United Kingdom

Christine M. Bachman

Roles: Project administration, Resources, Writing – review & editing

Affiliation: Global Health Labs, Inc, Bellevue, Washington, United States of America

https://orcid.org/0000-0002-8143-2782

Bryan Greenhouse

Roles: Conceptualization, Investigation, Writing – review & editing

Affiliation: Department of Medicine, University of California San Francisco, San Francisco, California, United States of America

Chris Drakeley

Roles: Conceptualization, Funding acquisition, Investigation, Writing – review & editing

Affiliation: Department of Infection Biology, London School of Hygiene and Tropical Medicine, London, United Kingdom

https://orcid.org/0000-0003-4863-075X

Phil L. Felgner

Roles: Conceptualization, Methodology, Resources, Writing – review & editing

Affiliation: Department of Physiology and Biophysics, University of California, Irvine, Irvine, CA, United States of America

Will Stone

Roles: Formal analysis, Investigation, Methodology, Supervision, Writing – review & editing

Affiliation: Department of Infection Biology, London School of Hygiene and Tropical Medicine, London, United Kingdom

https://orcid.org/0000-0002-6647-0166

Kevin K. A. Tetteh

Roles: Conceptualization, Formal analysis, Funding acquisition, Investigation, Resources, Supervision, Writing – original draft, Writing – review & editing

E-mail: [email protected]

Affiliation: Department of Infection Biology, London School of Hygiene and Tropical Medicine, London, United Kingdom

https://orcid.org/0000-0003-2733-0037

References

1. Corran P, Coleman P, Riley E, Drakeley C. Serology: a robust indicator of malaria transmission intensity? Trends Parasitol. 2007;23: 575–582. pmid:17988945

2. Drakeley CJ, Corran PH, Coleman PG, Tongren JE, McDonald SLR, Carneiro I, et al. Estimating medium- and long-term trends in malaria transmission by using serological markers of malaria exposure. Proc Natl Acad Sci U S A. 2005;102: 5108–5113. pmid:15792998

3. Stewart L, Gosling R, Griffin J, Gesase S, Campo J, Hashim R, et al. Rapid assessment of malaria transmission using age-specific sero-conversion rates. PloS One. 2009;4: e6083–e6083. pmid:19562032

4. Bousema T, Youssef RM, Cook J, Cox J, Alegana VA, Amran J, et al. Serologic markers for detecting malaria in areas of low endemicity, Somalia, 2008. Emerg Infect Dis. 2010;16: 392–399. pmid:20202412

5. Theisen M, Vuust J, Gottschau A, Jepsen S, Høgh B. Antigenicity and immunogenicity of recombinant glutamate-rich protein of Plasmodium falciparum expressed in Escherichia coli. Clin Diagn Lab Immunol. 1995;2: 30–34. pmid:7719909

6. Carvalho LH, Fontes CJ, Fernandes AA, Marinuzzi HC, Krettli AU. Cross-reactive cellular immune response to circumsporozoite proteins of Plasmodium vivax and P. falciparum in malaria-exposed individuals. Parasite Immunol. 1997;19: 47–59. pmid:9076807

7. Milek RL, DeVries AA, Roeffen WF, Stunnenberg H, Rottier PJ, Konings RN. Plasmodium falciparum: heterologous synthesis of the transmission-blocking vaccine candidate Pfs48/45 in recombinant vaccinia virus-infected cells. Exp Parasitol. 1998;90: 165–174. pmid:9769246

8. Kanunfre KA, Leoratti FMS, Hoffmann EHE, Durlacher RR, Ferreira AW, Moraes-Avila SL, et al. Differential recognition of Plasmodium falciparum merozoite surface protein 2 variants by antibodies from malaria patients in Brazil. Clin Diagn Lab Immunol. 2003;10: 973–976. pmid:12965937

9. Gardner MJ, Hall N, Fung E, White O, Berriman M, Hyman RW, et al. Genome sequence of the human malaria parasite Plasmodium falciparum. Nature. 2002;419: 498–511. pmid:12368864

10. Florens L, Washburn MP, Raine JD, Anthony RM, Grainger M, Haynes JD, et al. A proteomic view of the Plasmodium falciparum life cycle. Nature. 2002;419: 520–526. pmid:12368866

11. Burbelo PD, Ching KH, Bush ER, Han BL, Iadarola MJ. Antibody-profiling technologies for studying humoral responses to infectious agents. Expert Rev Vaccines. 2010;9: 567–578. pmid:20518713

12. Manzano-Román R, Dasilva N, Díez P, Díaz-Martín V, Pérez-Sánchez R, Orfao A, et al. Protein arrays as tool for studies at the host-pathogen interface. J Proteomics. 2013;94: 387–400. pmid:24140974

13. Adu B, Issahaque Q-A, Sarkodie-Addo T, Kumordjie S, Kyei-Baafour E, Sinclear CK, et al. Microscopic and Submicroscopic Asymptomatic Plasmodium falciparum Infections in Ghanaian Children and Protection against Febrile Malaria. Infect Immun. 2020;88. pmid:32719157

14. Ondigo BN, Hamre KES, Frosch AEP, Ayodo G, White MT, John CC. Antibody Profiles to P. falciparum Antigens Over Time Characterize Acute and Long-Term Malaria Exposure in an Area of Low and Unstable Transmission. Am J Trop Med Hyg. 2020. pmid:33124539

15. Davies DH, Liang XW, Hernandez JE, Randall A, Hirst S, Mu YX, et al. Profiling the humoral immune response to infection by using proteome microarrays: High-throughput vaccine and diagnostic antigen discovery. Proc Natl Acad Sci U S A. 2005;102: 547–552. pmid:15647345

16. Arevalo-Herrera M, Lopez-Perez M, Dotsey E, Jain A, Rubiano K, Felgner PL, et al. Antibody Profiling in Naive and Semi-immune Individuals Experimentally Challenged with Plasmodium vivax Sporozoites. PLoS Negl Trop Dis. 2016;10: e0004563. pmid:27014875

17. Crompton PD, Kayala MA, Traore B, Kayentao K, Ongoiba A, Weiss GE, et al. A prospective analysis of the Ab response to Plasmodium falciparum before and after a malaria season by protein microarray. Proc Natl Acad Sci U S A. 2010;107: 6958–6963. pmid:20351286

18. Sundaresh S, Randall A, Unal B, Petersen JM, Belisle JT, Hartley MG, et al. From protein microarrays to diagnostic antigen discovery: a study of the pathogen Francisella tularensis. Bioinformatics. 2007;23: I508–I518. pmid:17646338

19. Helb DA, Tetteh KKA, Felgner PL, Skinner J, Hubbard A, Arinaitwe E, et al. Novel serologic biomarkers provide accurate estimates of recent Plasmodium falciparum exposure for individuals and communities. Proc Natl Acad Sci. 2015;112: E4438–E4447. pmid:26216993

20. Stone WJR, Campo JJ, Ouédraogo AL, Meerstein-Kessel L, Morlais I, Da D, et al. Unravelling the immune signature of Plasmodium falciparum transmission-reducing immunity. Nat Commun. 2018;9: 558. pmid:29422648

21. van den Hoogen LL, Walk J, Oulton T, Reuling IJ, Reiling L, Beeson JG, et al. Antibody Responses to Antigenic Targets of Recent Exposure Are Associated With Low-Density Parasitemia in Controlled Human Plasmodium falciparum Infections. Front Microbiol. 2019;9. pmid:30700984

22. Sen S, Sanders EC, Gabriel KN, Miller BM, Isoda HM, Salcedo GS, et al. Antibody Response to a Nucleocapsid Epitope as a Marker for COVID-19 Disease Severity. BioRxiv Prepr Serv Biol. 2020.

23. Khan S, Nakajima R, Jain A, de Assis RR, Jasinskas A, Obiero JM, et al. Analysis of Serologic Cross-Reactivity Between Common Human Coronaviruses and SARS-CoV-2 Using Coronavirus Antigen Microarray. BioRxiv Prepr Serv Biol. 2020. pmid:32511324

24. Chakraborty S, Randall A, Vickers TJ, Molina D, Harro CD, DeNearing B, et al. Interrogation of a live-attenuated enterotoxigenic Escherichia coli vaccine highlights features unique to wild-type infection. NPJ Vaccines. 2019;4: 37. pmid:31482013

25. Murray CJ, Baliga R. Cell-free translation of peptides and proteins: from high throughput screening to clinical production. Curr Opin Chem Biol. 2013;17: 420–426. pmid:23499386

26. Dopp JL, Jo YR, Reuel NF. Methods to reduce variability in E. Coli-based cell-free protein expression experiments. Synth Syst Biotechnol. 2019;4: 204–211. pmid:31750411

27. Sridharan H, Piorino F, Styczynski MP. Systems biology-based analysis of cell-free systems. Curr Opin Biotechnol. 2022;75: 102703. pmid:35247659

28. Silverman AD, Karim AS, Jewett MC. Cell-free gene expression: an expanded repertoire of applications. Nat Rev Genet. 2020;21: 151–170. pmid:31780816

29. Venkatesh A, Jain A, Davies H, Felgner PL, Rathod PK, Patankar S, et al. Protein Arrays for the Identification of Seroreactive Protein Markers for Infectious Diseases. Methods Mol Biol Clifton NJ. 2021;2344: 139–150. pmid:34115357

30. Kobayashi T, Jain A, Liang L, Obiero JM, Hamapumbu H, Stevenson JC, et al. Distinct Antibody Signatures Associated with Different Malaria Transmission Intensities in Zambia and Zimbabwe. Blader IJ, editor. mSphere. 2019;4: e00061–19, /msphere/4/2/mSphere061-19.atom. pmid:30918058

31. Longley RJ, White MT, Takashima E, Brewster J, Morita M, Harbers M, et al. Development and validation of serological markers for detecting recent Plasmodium vivax infection. Nat Med. 2020;26: 741–749. pmid:32405064

32. Nagaoka H, Kanoi BN, Jinoka K, Morita M, Arumugam TU, Palacpac NMQ, et al. The N-Terminal Region of Plasmodium falciparum MSP10 Is a Target of Protective Antibodies in Malaria and Is Important for PfGAMA/PfMSP10 Interaction. Front Immunol. 2019;10: 2669. pmid:31824483

33. Reiling L, Boyle MJ, White MT, Wilson DW, Feng G, Weaver R, et al. Targets of complement-fixing antibodies in protective immunity against malaria in children. Nat Commun. 2019;10: 610. pmid:30723225

34. Ito D, Chen J-H, Takashima E, Hasegawa T, Otsuki H, Takeo S, et al. Identification of a Novel RAMA/RON3 Rhoptry Protein Complex in Plasmodium falciparum Merozoites. Front Cell Infect Microbiol. 2020;10: 605367. pmid:33537242

35. Kamuyu G, Tuju J, Kimathi R, Mwai K, Mburu J, Kibinge N, et al. KILchip v1.0: A Novel Plasmodium falciparum Merozoite Protein Microarray to Facilitate Malaria Vaccine Candidate Prioritization. Front Immunol. 2018;9: 2866. pmid:30619257

36. Berrow NS, Alderton D, Sainsbury S, Nettleship J, Assenberg R, Rahman N, et al. A versatile ligation-independent cloning method suitable for high-throughput expression screening applications. Nucleic Acids Res. 2007;35: e45. pmid:17317681

37. Alzari PM, Berglund H, Berrow NS, Blagova E, Busso D, Cambillau C, et al. Implementation of semi-automated cloning and prokaryotic expression screening: the impact of SPINE. Acta Crystallogr D Biol Crystallogr. 2006;62: 1103–1113. pmid:17001088

38. Nettleship JE, Rada H, Owens RJ. Overview of a High-Throughput Pipeline for Streamlining the Production of Recombinant Proteins. Methods Mol Biol Clifton NJ. 2019;2025: 33–49. pmid:31267447

39. Baneyx F, Mujacic M. Recombinant protein folding and misfolding in Escherichia coli. Nat Biotechnol. 2004;22: 1399–1408. pmid:15529165

40. Ban B, Sharma M, Shetty J. Optimization of Methods for the Production and Refolding of Biologically Active Disulfide Bond-Rich Antibody Fragments in Microbial Hosts. Antibodies Basel Switz. 2020;9. pmid:32764309

41. Francis DM, Page R. Strategies to Optimize Protein Expression in E. coli. Curr Protoc Protein Sci. 2010;61. pmid:20814932

42. Vedadi M, Lew J, Artz J, Amani M, Zhao Y, Dong A, et al. Genome-scale protein expression and structural biology of Plasmodium falciparum and related Apicomplexan organisms. Mol Biochem Parasitol. 2007;151: 100–110. pmid:17125854

43. Mehlin C, Boni E, Buckner FS, Engel L, Feist T, Gelb MH, et al. Heterologous expression of proteins from Plasmodium falciparum: results from 1000 genes. Mol Biochem Parasitol. 2006;148: 144–160. pmid:16644028

44. Kamya MR, Arinaitwe E, Wanzira H, Katureebe A, Barusya C, Kigozi SP, et al. Malaria transmission, infection, and disease at three sites with varied transmission intensity in Uganda: implications for malaria control. Am J Trop Med Hyg. 2015;92: 903–912. pmid:25778501

45. Burghaus PA, Holder AA. Expression of the 19-kilodalton carboxy-terminal fragment of the Plasmodium falciparum merozoite surface protein-1 in Escherichia coli as a correctly folded protein. Mol Biochem Parasitol. 1994;64: 165–169. pmid:8078519

46. Polley SD, Tetteh KKA, Cavanagh DR, Pearce RJ, Lloyd JM, Bojang KA, et al. Repeat sequences in block 2 of Plasmodium falciparum merozoite surface protein 1 are targets of antibodies associated with protection from malaria. Infect Immun. 2003;71: 1833–1842. pmid:12654798

47. Spielmann T, Fergusen DJP, Beck H-P. etramps, a new Plasmodium falciparum gene family coding for developmentally regulated and highly charged membrane proteins located at the parasite-host cell interface. Mol Biol Cell. 2003;14: 1529–1544. pmid:12686607

48. Richards JS, Stanisic DI, Fowkes FJI, Tavul L, Dabod E, Thompson JK, et al. Association between naturally acquired antibodies to erythrocyte-binding antigens of Plasmodium falciparum and protection from malaria and high-density parasitemia. Clin Infect Dis Off Publ Infect Dis Soc Am. 2010;51: e50–60. pmid:20843207

49. Wang L, Menting JG, Black CG, Stowers A, Kaslow DC, Hoffman SL, et al. Differences in epitope recognition, isotype and titer of antisera to Plasmodium falciparum merozoite surface protein 4 raised by different modes of DNA or protein immunization. Vaccine. 2000;19: 816–824. pmid:11115704

50. Black CG, Barnwell JW, Huber CS, Galinski MR, Coppel RL. The Plasmodium vivax homologues of merozoite surface proteins 4 and 5 from Plasmodium falciparum are expressed at different locations in the merozoite. Mol Biochem Parasitol. 2002;120: 215–224. pmid:11897127

51. Pachebat JA, Ling IT, Grainger M, Trucco C, Howell S, Fernandez-Reyes D, et al. The 22 kDa component of the protein complex on the surface of Plasmodium falciparum merozoites is derived from a larger precursor, merozoite surface protein 7. Mol Biochem Parasitol. 2001;117: 83–89. pmid:11551634

52. Hinds L, Green JL, Knuepfer E, Grainger M, Holder AA. Novel putative glycosylphosphatidylinositol-anchored micronemal antigen of Plasmodium falciparum that binds to erythrocytes. Eukaryot Cell. 2009;8: 1869–1879. pmid:19820120

53. Collins CR, Withers-Martinez C, Bentley GA, Batchelor AH, Thomas AW, Blackman MJ. Fine mapping of an epitope recognized by an invasion-inhibitory monoclonal antibody on the malaria vaccine candidate apical membrane antigen 1. J Biol Chem. 2007;282: 7431–7441. pmid:17192270

54. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43: e47–e47. pmid:25605792

55. Collins CR, Withers-Martinez C, Hackett F, Blackman MJ. An inhibitory antibody blocks interactions between components of the malarial invasion machinery. PLoS Pathog. 2009;5: e1000273. pmid:19165323

56. Richards JS, Arumugam TU, Reiling L, Healer J, Hodder AN, Fowkes FJI, et al. Identification and Prioritization of Merozoite Antigens as Targets of Protective Human Immunity to Plasmodium falciparum Malaria for Vaccine and Biomarker Development. J Immunol. 2013;191: 795. pmid:23776179

57. Bailey JA, Berry AA, Travassos MA, Ouattara A, Boudova S, Dotsey EY, et al. Microarray analyses reveal strain-specific antibody responses to Plasmodium falciparum apical membrane antigen 1 variants following natural infection and vaccination. Sci Rep. 2020;10: 3952. pmid:32127565

58. Tessema SK, Nakajima R, Jasinskas A, Monk SL, Lekieffre L, Lin E, et al. Protective Immunity against Severe Malaria in Children Is Associated with a Limited Repertoire of Antibodies to Conserved PfEMP1 Variants. Cell Host Microbe. 2019;26: 579–590.e5. pmid:31726028

59. Proietti C, Krause L, Trieu A, Dodoo D, Gyan B, Koram KA, et al. Immune Signature Against Plasmodium falciparum Antigens Predicts Clinical Immunity in Distinct Malaria Endemic Communities. Mol Cell Proteomics MCP. 2020;19: 101–113. pmid:31658979

60. Stucke EM, Niangaly A, Berry AA, Bailey JA, Coulibaly D, Ouattara A, et al. Serologic responses to the PfEMP1 DBL-CIDR head structure may be a better indicator of malaria exposure than those to the DBL-α tag. Malar J. 2019;18: 273. pmid:31409360

61. Venkatesh A, Jain A, Davies H, Periera L, Maki JN, Gomes E, et al. Hospital-derived antibody profiles of malaria patients in Southwest India. Malar J. 2019;18: 138. pmid:30995911

62. Friedman-Klabanoff DJ, Travassos MA, Ifeonu OO, Agrawal S, Ouattara A, Pike A, et al. Epitope-specific antibody responses to a Plasmodium falciparum subunit vaccine target in a malaria-endemic population. J Infect Dis. 2020. pmid:32992328

63. Wu L, Mwesigwa J, Affara M, Bah M, Correa S, Hall T, et al. Sero-epidemiological evaluation of malaria transmission in The Gambia before and after mass drug administration. BMC Med. 2020;18: 331. pmid:33183292

64. Labadie-Bracho MY, van Genderen FT, Adhin MR. Malaria serology data from the Guiana shield: first insight in IgG antibody responses to Plasmodium falciparum, Plasmodium vivax and Plasmodium malariae antigens in Suriname. Malar J. 2020;19: 360. pmid:33032606

65. Wu L, Mwesigwa J, Affara M, Bah M, Correa S, Hall T, et al. Antibody responses to a suite of novel serological markers for malaria surveillance demonstrate strong correlation with clinical and parasitological infection across seasons and transmission settings in The Gambia. BMC Med. 2020;18: 304. pmid:32972398

66. Steinhardt L, Ravaoarisoa E, Wiegand R, Harimanana A, Hedje J, Cotte AH, et al. School-based sero-surveys to assess the validity of using routine health facility data to target malaria interventions in the Central Highlands of Madagascar. J Infect Dis. 2020. pmid:32761176

67. Lu A, Cote O, Dimitrova SD, Cooley G, Alamgir A, Uzzaman MS, et al. Screening for malaria antigen and anti-malarial IgG antibody in forcibly-displaced Myanmar nationals: Cox’s Bazar district, Bangladesh, 2018. Malar J. 2020;19: 130. pmid:32228699

68. Surendra H, Supargiyono , Ahmad RA, Kusumasari RA, Rahayujati TB, Damayanti SY, et al. Using health facility-based serological surveillance to predict receptive areas at risk of malaria outbreaks in elimination areas. BMC Med. 2020;18: 9. pmid:31987052

69. Müller-Sienerth N, Shilts J, Kadir KA, Yman V, Homann MV, Asghar M, et al. A panel of recombinant proteins from human-infective Plasmodium species for serological surveillance. Malar J. 2020;19: 31. pmid:31952523

70. Assefa A, Ali Ahmed A, Deressa W, Sime H, Mohammed H, Kebede A, et al. Multiplex serology demonstrate cumulative prevalence and spatial distribution of malaria in Ethiopia. Malar J. 2019;18: 246. pmid:31331340

71. Iskakova MB, Szaflarski W, Dreyfus M, Remme J, Nierhaus KH. Troubleshooting coupled in vitro transcription–translation system derived from Escherichia coli cells: synthesis of high-yield fully active proteins. Nucleic Acids Res. 2006;34: e135–e135. pmid:17038334

72. Caschera F, Noireaux V. Synthesis of 2.3 mg/ml of protein with an all Escherichia coli cell-free transcription–translation system. Biochimie. 2014;99: 162–168. pmid:24326247

73. Chizzolini F, Forlin M, Yeh Martín N, Berloffa G, Cecchi D, Mansy SS. Cell-Free Translation Is More Variable than Transcription. ACS Synth Biol. 2017;6: 638–647. pmid:28100049

74. Morita M, Takashima E, Ito D, Miura K, Thongkukiatkul A, Diouf A, et al. Immunoscreening of Plasmodium falciparum proteins expressed in a wheat germ cell-free system reveals a novel malaria vaccine candidate. Sci Rep. 2017;7: 46086. pmid:28378857

75. Kanoi BN, Takashima E, Morita M, White MT, Palacpac NMQ, Ntege EH, et al. Antibody profiles to wheat germ cell-free system synthesized Plasmodium falciparum proteins correlate with protection from symptomatic malaria in Uganda. Vaccine. 2017;35: 873–881. pmid:28089547

76. Nettleship JE, Assenberg R, Diprose JM, Rahman-Huq N, Owens RJ. Recent advances in the production of proteins in insect and mammalian cells for structural biology. J Struct Biol. 2010;172: 55–65. pmid:20153433

77. Konczal J, Gray CH. Streamlining workflow and automation to accelerate laboratory scale protein production. Protein Expr Purif. 2017;133: 160–169. pmid:28330825