Potential of High-Affinity, Slow Off-Rate Modified Aptamer

48 Pages·2017·1.35 MB·English

JCM Accepted Manuscript Posted Online 9 August 2017
J. Clin. Microbiol. doi:10.1128/JCM.00469-17
Copyright © 2017 Russell et al.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.
1 JCM00469-17R1 for submission to J Clin Microbiol revised 07-08-17
2 (Accompanying manuscript with De Groote MA et al.)
3
4 Potential of High-Affinity, Slow Off-Rate Modified Aptamer (SOMAmer)
5 Reagents for Mycobacterium tuberculosis Proteins as Tools for Infection
D
o
w
6 Models and Diagnostic Applications n
lo
a
7 d
e
d
8 fr
o
m
9 Running Title: M. tuberculosis SOMAmer reagents
h
t
t
p
10 :
/
/
jc
11 Theresa Russell,a Louis S. Green,a Taylor Rice,a Nicole A. Kruh-Garcia,b Karen Dobos,b Mary m
.
a
s
12 A. DeGroote,a,b Thomas Hraha,a David G. Sterling,a Nebojsa Janjic,a Urs A. Ochsnera m
.
o
13 rg
/

o
14 SomaLogic, Inc., Boulder, Colorado, USAa, Department of Microbiology, Immunology & n

A
p
15 Pathology, Colorado State University, Fort Collins, CO, USAb ril
2
,
16 2
0
1
17 Address correspondence to Urs Ochsner, [email protected] 9

b
y

g
u
e
s
t
1
18 ABSTRACT (250 words allowed – now 247)
19 Direct pathogen detection in blood to diagnose active tuberculosis (TB) has been difficult, due to
20 low levels of circulating antigens or due to the lack of specific, high-affinity binding reagents
21 and reliable assays with adequate sensitivity. We sought to determine whether slow off-rate
22 modified aptamer (SOMAmer) reagents with sub-nanomolar affinity for M. tuberculosis (Mtb)
D
o
23 proteins (antigen 85A, 85B, 85C, GroES, GroEL2, DnaK, CFP10, KAD, CFP2, RplL and Tpx) w
n
24 could be useful to diagnose tuberculosis. When incorporated into the multiplexed, array-based lo
a
d
e
25 proteomic SOMAscan assay, limits of detection reached the sub-picomolar range in 40% serum. d

f
r
o
26 Binding to native Mtb proteins was confirmed by using Mtb culture filtrate proteins, fractions m

h
27 from infected macrophages, and via affinity-capture assays and subsequent mass spectrometry. tt
p
:
/
/
28 Comparison of serum from culture-positive pulmonary TB patients and TB-suspects jc
m
.
29 systematically ruled out for TB revealed small but statistically significant (p<0.0001) differences a
s
m
30 in the median Mtb signals and in specific pathogen markers such as antigen 85B (A85B). .o
r
g
/
31 Samples where many Mtb aptamers produced high signals were rare exceptions. In concentrated, o
n

A
32 protein-normalized urine from TB patients and non-TB controls, the CFP10 (EsxB) SOMAmer
p
r
33 yielded the most significant differential signals (p<0.0276), particularly in TB patients with HIV il 2
,

2
34 co-infection. In conclusion, direct Mtb antigen detection proved difficult even with a sensitive 0
1
9

35 method such as SOMAscan, likely due to their very low, sub-picomolar abundance. The b
y

g
36 observed differences between cases and controls had limited diagnostic utility in serum and u
e
s
t
37 urine, but further evaluation of Mtb SOMAmers using other platforms and sample types is
38 warranted.
2
39 (INTRODUCTION)
40 Tuberculosis (TB) remains a major global health problem and in 2015 had the highest mortality
41 of any infectious disease worldwide. While there has been a steady yet slow decline in new TB
42 cases at a rate of 2% per year recently, incidence remains high in Africa, particularly in sub-
43 Saharan Africa, Asia, the Western Pacific Region, and in Central and South America, totaling
D
o
44 10.4 million new TB cases in the world in 2015 (1). TB mortality rate has decreased by almost w
n
45 half since 1990 but there are still over a million deaths per year with about one-fourth of these lo
a
d
e
46 deaths occurring among people with HIV. d

f
r
o
47 An estimated 4.3 million new TB cases a year remain undiagnosed (1, 2). Since pulmonary TB m

h
48 represents the majority of TB cases and is transmitted via aerosols from people with active tt
p
:
/
/
49 pulmonary disease, a high diagnostic priority is to determine those with active TB to enable rapid jc
m
.
50 treatment and reduce disease transmission. High-priority diagnostic needs for which specific a
s
m
51 target product profiles (TPPs) have been defined include a non-sputum-based biomarker test for .o
r
g
/
52 all forms of TB, and a simple, low cost triage test for use by first-contact care providers as a rule- o
n

A
53 out test (3-5). Traditional methods for pathogen detection are culture and/or staining microscopy
p
r
54 of the Mtb bacilli. Several classes of Mtb-specific pathogen products are potential candidates for il 2
,

2
55 new diagnostic methods, including lipoarabinomannan (LAM), metabolites (including lipids, 0
1
9

56 sugars, and volatile organic compounds in breath), nucleic acids and peptides and proteins (6, 7). b
y

g
57 Sputum smear microscopy for TB diagnosis is still widely used around the world and is fast and u
e
s
t
58 inexpensive, though suboptimal in children and in people with HIV (8, 9). Culture-based
59 diagnostic methods are more sensitive, but require several weeks to obtain results (10), which
60 causes a delay in therapy. The tuberculin skin test (TST) and the interferon-gamma release assay
61 (IGRA) measure the immune response to Mtb antigens and thus do not distinguish active TB
3
62 from latent disease or a previously cleared infection (11, 12). Gene-Xpert MTB/RIF is a rapid
63 sputum molecular diagnostic test which has been rolled out in many countries and performs very
64 well, except perhaps in smear-negative TB and pediatric TB cases (13-17). Gene-Xpert has
65 transformed TB diagnostics, though it requires complex and expensive cartridges and a reliable
66 power source. Alternative rapid, accurate tests for point-of-care TB diagnostics using non-
D
o
67 sputum based samples and new or improved technologies are critically needed (6). w
n
68 Our proteomic technology is based on affinity-binding reagents and a technology platform lo
a
d
e
69 targeting proteins that are intact or at a minimum harbor the native structural epitopes used for d

f
r
o
70 selection of the binding reagents. SOMAmers are a new class of synthetic reagents, in some m

h
71 ways similar to monoclonal antibodies, and are used in proteomic applications where high tt
p
:
/
/
72 sensitivity and specificity is needed. Advantages of SOMAmers over antibodies include higher jc
m
.
73 multiplexing capabilities due to low cross-reactivity and universal assay conditions, chemical a
s
m
74 stability to heat, drying, and solvents, reversible renaturation, ease of reagent manufacturing, .o
r
g
/
75 consistent lot-to-lot performance, and lower cost. Modified aptamers are the basis for the o
n

A
76 SOMAscan multiplex proteomic assay that measures thousands of proteins simultaneously with
p
r
77 high precision (<5% CV) in a small sample volume of <150 µl. The overall dynamic range of il 2
,

2
78 the assay is roughly 8 logs, with a median lower limit of detection of 40 femtomolar (18, 19). 0
1
9

79 We initiated a broad TB biomarker discovery effort using serum and urine samples from initial b
y

g
80 TB-suspects that had confirmed TB or were ruled out for TB (non-TB, NTB) based on u
e
s
t
81 protocolized culture and and systematic follow-up. In this report we focus on the generation of
82 specific high-affinity binding reagents to Mtb pathogen-derived proteins and evaluate their utility
83 for direct Mtb antigen detection. A separate, accompanying report by De Groote et al. describes
4
84 the identification of host response markers and the performance of a TB-specific biosignature
85 entirely based on host markers (20).
86 A multitude of Mtb proteins may be useful as potential diagnostic TB targets, given the observed
87 immunologic responses in TB patients stemming from serology studies or infection models (21-
88 24). We chose 18 mycobacterial proteins for Mtb SOMAmer development via Systematic
D
o
89 Evolution of Ligands by Exponential Enrichment (SELEX) (19, 25). Mtb protein targets w
n
90 included extracellular, cell-surface-associated, and intracellular factors that may be circulating in lo
a
d
e
91 TB patients and may be detectable with a highly sensitive and specific assay. FbpA, FbpB, FpbC d

f
r
o
92 form the antigen-85 complex and are highly abundant mycolyltransferases essential for cell wall m

h
93 synthesis and are also major secretory antigens (26, 27). ESAT-6 (EsxA) and CFP10 (EsxB) tt
p
:
/
/
94 represent excellent diagnostic targets since they are absent in M. bovis BCG (28). PstS1 is a jc
m
.
95 Mycobacteria-specific lipoprotein phosphate transporter (29). MPT64 and MPT51 are major a
s
m
96 extracellular antigens and have once been considered to improve the BCG vaccine (30, 31). α– .o
r
g
/
97 Crystalline (Acr, HspX) is an abundant inner membrane protein induced by microaerobic and o
n

A
98 anoxic conditions and plays a role in the long-term viability during latent infection (32). CFP30
p
r
99 and MTB12 (CFP-2) are found in culture supernatants and are known antigens during infection il 2
,

2
100 (33, 34). Other targets described as potential diagnostic markers were GroES (CH10), GroEL2 0
1
9

101 (CH602), DnaK, 50S ribosomal protein RplL, Adk (KAD), MasZ, and Tpx (35-39). These b
y

g
102 intracellular proteins are produced constitutively and at high levels, though these factors are less u
e
s
t
103 specific for Mtb since closely related proteins are found in non-tuberculous mycobacteria (NTM)
104 and other Actinobacteria such as Nocardia and Streptomyces.
105 Here we describe the characterization of the pathogen-specific aptamers and the identification of
106 those with the greatest sensitivity and specificity against recombinant Mtb proteins, natively
5
107 expressed Mtb culture filtrate proteins, and fractions from Mtb-infected human macrophages.
108 Finally, the top performing Mtb aptamers were then examined as potential diagnostic tools in
109 well-curated clinical samples.
110
111 MATERIALS AND METHODS
D
o
112 Buffers and reagents w
n
113 Serum Sample Diluent (SSD) is composed of 40 mM HEPES, 61 mM NaCl, 3 mM KCl, 5 mM lo
a
d
e
114 MgCl2, 0.6 mM EGTA, 1.2 mM benzamidine, 0.02 mM Z-block (see below), 0.05% (v/v) d

f
r
o
115 Tween 20 and adjusted to pH 7.5 with NaOH. Buffer SB18T is composed of 40 mM HEPES, m

h
116 120 mM NaCl, 5 mM KCl, 5 mM MgCl2, and 0.05% (v/v) Tween 20 adjusted to pH 7.5 with tt
p
:
/
/
117 NaOH. Buffer SB17T is SB18T supplemented with 1mM EDTA. jc
m
.
118 Z-block is a SOMAmer mimic and was used at 0.1-20 µM to block non-specific interactions of a
s
m
119 reagents with proteins and was prepared at SomaLogic. Dextran sulfate (Sigma) was another .o
r
g
/
120 polyanionic competitor and was prepared as 10 mM solution in SB17T. o
n

A
121
p
r
122 M. tuberculosis proteins and culture filtrate proteins il 2
,

2
123 Eight Mtb genes were PCR-amplified from genomic DNA of strain H37Rv (BEI Resources, NR- 0
1
9

124 14865) using KOD XL DNA polymerase (EMD Millipore) and ligated into pET-51b vector b
y

g
125 (EMD Millipore) using restriction sites contained in the primer sequences (Table S1). For 10 u
e
s
t
126 additional targets, pET-15b or pET-23a based expression vectors were available from BEI
127 Resources (Table S1). All plasmids were sequenced (SeqWright) to verify the identity of the
128 cloned genes and their proper in-frame fusion with the vector-encoded tag(s). E. coli Rosetta
129 (EMD Millipore) was transformed with these plasmids and grown in 400 ml of LB medium for
6
130 the production of bacterial proteins as previously described (40, 41). In brief, expression was
131 induced with 0.5 mM IPTG during mid-exponential growth (OD600=0.3-0.7), and cultures were
132 shaken for 4-16 h at 30-35°C. After cell lysis with BugBuster/Benzonase reagent (EMD
133 Millipore), the recombinant His-tagged proteins were purified from the soluble fraction via
134 affinity chromatography on Ni-NTA agarose. Proteins that also contained a Strep-tag were
D
135 further purified using StrepTactin® Superflow™ agarose (EMD Millipore). All 18 Mtb proteins ow
n
lo
136 were obtained in sufficient amounts and purity for use in SELEX and subsequent binding and a
d
e
137 spiking assays (Table S1). d

f
r
o
138 Culture filtrate proteins (CFP) from M. tuberculosis strains H37Rv, CDC1551, and HN878 were m

h
139 obtained from BEI Resources (Manassas, VA) and their preparation was described previously tt
p
:
/
/
140 (42). The strains represent a lab-adapted (H37Rv) and a clinical isolate (CDC1551) from the jc
m
.
141 Euro-American lineage and a clinical isolate (HN878) from the East-Asian lineage (43). a
s
m
142 Recovery of CFP for these strains by this method ranged from 11-12 µg/mg of wet cell paste, .o
r
g
/
143 with roughly 8 x 108 cells/mg of wet cell paste. o
n

A
144 p
r
il
145 SOMAmer selection and synthesis 2
,

2
0
146 Seven different modified nucleotide libraries were used for SELEX, including 2NapdU, 2NEdU, 1
9

b
147 BndU, NapdU, PEdU, PPdU, and TrpdU (44), using methods previously described (19, 40). y

g
u
148 Special effort was devoted to generating Mtb SOMAmer reagents with minimal non-specific e
s
t
149 background in serum, by applying counter-selection with serum competitor buffer (SCB: 1 µM
150 casein, 1 µM prothrombin, 7.5% serum) and with protein competitor buffer (PCB: 1 µM casein,
151 1 µM prothrombin, 1.5 µM albumin) in alternating SELEX rounds. For some selections of
152 antigen-85 (A85) SOMAmers, complexes with fibronectin (FN) were pre-formed by mixing the
7
153 two proteins together in 1:1 stoichiometric ratios, and including a 4-fold molar excess of
154 untagged, free fibronectin for counter-selection during SELEX. For all targets, increasingly
155 longer (up to 90 min) kinetic challenges with 10 mM dextran sulfate were performed to favor
156 slow off-rates. SOMAmer–target complexes were partitioned with paramagnetic Dynabeads®
157 His-Tag beads (ThermoFisher) that bind the His-tag on the recombinant proteins, and selected
D
o
158 aptamers were amplified using KOD XL DNA polymerase. Aptamer pools with good affinity w
n
159 (Kd ≤10 nM) were deep-sequenced, and enriched sequences were prepared synthetically as 50- loa
d
e
160 mers via standard phosphoramidite chemistry, incorporating a 5’PBDC (photocleavable biotin, D d

f
r
o
161 spacer, cyanine 3) and an inverted dT nucleotide at the 3’ end (3’idT) for added stability to 3’ to m

h
162 5’ exonucleases. After initial characterization, the best SOMAmers were prepared at larger scale tt
p
:
/
/
163 (1 µmol) and HPLC-purified. jc
m
.
164 a
s
m
165 SOMAmer characterization .o
r
g
/
166 SOMAmers were refolded by denaturation for 5 min at 95°C, followed by cooling to room o
n

167 temperature over a 1015 min period. For equilibrium solution binding assays, 32P-radiolabeled Ap
r
il
168 aptamers (10-50 pM) were incubated for 2 h at 37°C with serially diluted proteins (0.001100 2
,

2
0
169 nM). Zorbax PSM-300A (Agilent Technologies) resin or Dynabeads® His-Tag beads 1
9

b
170 (ThermoFisher) were used for partitioning of the complexes. Aptamer affinities were calculated y

g
u
171 from the resulting binding curves and expressed as equilibrium binding constants (Kd’s), along e
s
t
172 with the maximum bound fraction (binding plateau) which is influenced by the retention
173 efficiency of the targetSOMAmer complexes on Zorbax, the fraction of binding-competent
174 aptamers, and traces of unincorporated radiolabel. Apparent Kd (Kd, app) values were determined
8
175 in the proteomic SOMAscan assay (described below), where each aptamer was present at higher
176 concentration (500 pM) and equilibration time was longer (3.5 h).
177
178 Affinity capture and LC-MS/MS protein identification
179 Affinity-capture assays (pull-downs) were performed as previously reported (19) with
D
o
180 modifications. Agarose-streptavidin beads, with 10 pmoles of pre-immobilized SOMAmers were w
n
181 pre-blocked (SB17, 1% BSA, 2 mg/ml sheared salmon sperm DNA) for 1 h at 850 rpm and room lo
a
d
e
182 temperature. After washing, unoccupied streptavidin was blocked with biotin (SB17T, 10 mM d

f
r
o
183 biotin). Recombinant proteins (25 nM) and CFPs (20 µg) were diluted in serum (40%, final) or m

h
184 buffer and SSD (1X, final). Subject serum samples were likewise diluted to 40% in 1X SSD. tt
p
:
/
/
185 Samples were added to the SOMAmer beads and incubated for 3 h at 850 rpm and 37°C. jc
m
.
186 Unbound or non-specifically retained proteins were removed by washing with 10 mM dextran a
s
m
187 sulfate and three washes with SB17T (5 min, 850 rpm). Captured targets were labeled with 200 .o
r
g
/
188 µM EZ Link NHS-PEO4-biotin (Pierce) and 50 µM NHS-Alexa Fluor® Dye647 (Invitrogen) for o
n

A
189 5 min. After quenching and washing, complexes were released from the beads via photo- p
r
il
190 cleavage of the biotin-SOMAmer linkage (UV exposure, 20 minutes, 850 rpm) and then captured 2
,

2
0
191 on DynaBeads® MyOne™ streptavidin beads (Invitrogen). Traces of non-biotinylated protein, 1
9

b
192 non-specifically retained reagents, and free SOMAmers were removed by sequential washing y

g
u
193 with 30% glycerol (in SB17T) and SB17T. Protein-SOMAmer complexes were eluted from the e
s
t
194 beads by boiling in 1X SDS-PAGE sample buffer (Invitrogen), analyzed by SDS-PAGE, and the
195 fluorescence imaged using cyanine-3 (aptamer reagents) and cyanine-5 (Alexa Fluor® Dye647-
196 labeled proteins) settings.
9
197 Affinity-capture eluates were also evaluated by LC-MS/MS and tandem mass spectra were
198 collected and processed at MSBioworks (Ann Arbor, MI). For detergent removal, eluates were
199 loaded onto a 10% Bis-Tris SDS-PAGE gel (Novex, Invitrogen) and separated approximately 1
200 cm. The gel was stained with Coommassie, the entire mobility region was excised into one
201 segment, and in-gel trypsin digestion was performed. Digests were analyzed by nano LC/MS/MS
D
o
202 with a Waters NanoAcquity HPLC system interfaced to a ThermoFisher Q Exactive. Peptides w
n
203 were loaded on a trapping column and eluted over a 75 μm analytical column at 350 nL/min; lo
a
d
e
204 both columns were packed with Jupiter Luna C18 resin (Phenomenex). A 1 h gradient was d

f
r
o
205 employed. The mass spectrometer was operated in data-dependent mode, with MS and MS/MS m

h
206 performed in the Orbitrap at 70,000 FWHM resolution and 17,500 FWHM resolution, tt
p
:
/
/
207 respectively. The fifteen most abundant ions were selected for MS/MS. Data Processing Data jc
m
.
208 were searched using a local copy of Mascot with the following parameters: Enzyme: Trypsin, a
s
m
209 Database: SwissProt Human + MYCTU sequences (forward and reverse appended with common .o
r
g
/
210 contaminants), Fixed modification: Carbamidomethyl (C), Variable modifications: Oxidation o
n

A
211 (M), Acetyl (Protein N-term), Deamidation (NQ), Pyro-Glu (N-term Q), Mass values:
p
r
212 Monoisotopic Peptide Mass Tolerance: 10 ppm, Fragment Mass Tolerance: 0.02 Da, Max il 2
,

2
213 Missed Cleavages: 2. Mascot DAT files were parsed into the Scaffold software for validation, 0
1
9

214 filtering and to create a nonredundant list per sample. Data were filtered 1% protein and peptide b
y

g
215 FDR and requiring at least two unique peptides per protein. u
e
s
t
216
217 SOMAscan and SOMApanel arrays for proteomic analysis
218 Proteomic analysis was performed as previously described (19, 45), except where assay
219 conditions were modified or samples underwent pre-treatment as indicated. Briefly, serum
10